Clinical Psychology Review, Vol. 11, pp. 21 l-230, 1991 0272-7358191 $3.00 + .

oo
Printed in the USA. All rights reserved. Copyright 0 1991 Pergamon Press plc

COGNITIVE-BEHAVIORAL
INTERVENTIONS IN THE
PSYCHOANALYTIC PSYCHOTHERAPY OF
BORDERLINE PERSONALITY DISORDERS’

Drew Westen

University of Michigan

ABSTRACT. A pathology of affect regulation and of object relations (the cognitive and
affective processes that mediate interpersonal functioning in close relationships) are two
features that define much of borderline pathology. Cognitive-behavioral interventions that
target self-regulation and social-cognitive processes (such as perspective-taking and attribu-
tion) can be usefully integrated into psychodynamically oriented treatment of borderline
patients. Strictly psychodynamic approaches tend to be limited by factors such as a lack of
attention to processes by which conscious insight and coping strategies can be transformed into
automatic or adaptive unconscious procedures in these patients. Cognitive approaches tend to
be limited by rationalistic assumptions about motivation and an underestimation of the
complexities of cognitive-affective interactions and unconscious processes.

If researchers and clinicians who study borderline personality disorders (BPD)

agree on one thing, it is that treatment of these patients is extremely difficult and
frequently unsuccessful (see Waldinger, 1987). Waldinger and Gunderson (1984)
found that, for borderline patients of experienced clinicians, only half continued
in treatment beyond six months, only one-third completed treatment, and the
majority terminated against their therapist’s advice.2 Pharmacotherapy for bor-
derline patients can sometimes be helpful, especially when used adjunctively with
psychotherapy. However, medication typically produces modest results, some-
times generates unwanted effects [such as behavioral dyscontrol or impulsivity

‘A version of this article was presented at the Third Annual Convention of the Society for
the Exploration of Psychotherapy Integration, Chicago, IL, May, 1987.
‘What is particularly sobering about these findings is that these patients were apparently
relatively high-functioning borderlines, treated in private practice as outpatients, and that
their therapists were largly clinicians who have made important contributions to the
borderline literature.
Correspondence should be addressed to Drew Westen, Department of Psychology,
University of Michigan, Ann Arbor, MI 48109-1346.

211
212 D. Westen

when benzodiazepines are used to treat anxiety symptoms (Gardner & Cowdry,
1985)], and is often not as efficacious with these patients as in nonborderline
patients with similar Axis I pathology (Cowdry & Gardner, 1988; Goldberg,
Schultz, Resnick, Hamer, 8c Friedel, 1986; Gunderson, 1986; Liebowitz, Stone, AC
Turkat, 1986; Pilkonis SC Frank, 1988; Soloff et al., 1986).
The tendency of borderline patients to draw particular countertransference
constellations should lead to special circumspection in altering therapeutic tech-
nique or adopting new treatment strategies. For example, BPDs tend to elicit
wishes to save and wishes to kill, complementary to their wishes to be magically
protected and to destroy. These may evoke therapist responses such as suffocating
pseudo-nurturance that denies the patient’s aggression and sense of entitlement;
hostile interpretations or “showdowns,” disguised as necessary confrontations; or
aggressive use of multiple medications with limited efficacy, used to quell the
therapist’s anxiety or feelings of helplessness or to punish or subdue the patient.
Limita~ons of current approaches to the treatment of borderline disorders,
however, clearly suggest the need for continued consideration of other thera-
peutic interventions that may prove beneficial, particularly in light of the dearth of
empirical evidence supporting the efficacy of traditional treatments.
Behavioral and cognitive-behavioral paradigms have historically lacked the
theoretical tools to grapple with the complexities of personality disorders. Hunt’s
(1977) work on behavioral perspectives in the treatment of borderline patients is a
highly perspicacious exception, although Hunt, like others (Lehrer, Schiff, & Kris,
1971), focuses on the integration of operant conditioning principles into psy-
chodynami~ally oriented inpatient programs. The high rate of comorbidity of
BPD and mood disorders (Gunderson SC Elliot, 1985; Kroll & Ogata, 1987)
suggests that many patients treated by cognitive therapists for depression may
have concurrent borderline pathology, although this is seldom discussed in the
cognitive-behavioral literature (see Freeman & Leaf, 1989). Cognitive-behavioral
theories lack the conceptual scaffolding to distinguish depression as found in
borderlines from nonborderline depression, which may have very significant
implications for treatment. 3 Recent research comparing borderline with depressed
subjects using the Blatt et al. (1982) Depressive Experiences Questionnaire with
both adolescent and adult samples has documented the psychoanalytic view (see
Gunderson, 1984; Kernberg, 1975; Masterson, 1976) that BPDs suffer from a
qualitatively distinct kind of depression, characterized by emptiness, loneliness,
desperation in relation to attachment figures, and labile, diffuse negative affectiv-
ity (Wixom, 1988; Westen, Moses, Silk, et al., 1989).
Recently, borderline pathology has begun receiving more attention from cogni-
tive-behavioral clinicians. Linehan (1987a, 1987b, 1987c) has been developing a
cognitive-behavioral model for understanding and treating borderline patients,

‘Astute cognitive-behavioral clinicians are well aware of the diagnostic and prognostic
importance of the distinction. In speaking with several cognitive-behavioral colleagues
nationally, it appears that they typically do not treat borderline patients, do not expect
much therapeutic success with borderline patients who are depressed, or adjust their
treatments to include a substantial focus on patient-therapist interactions and expect the
treatment to last a period of years rather than months. Recent efforts to address personality
disorders from a cognitive perspective (Beck, Freeman et al., 1990; Freeman & Leaf, 1989)
are very preliminary, largely translating DSM III-R criteria into schema language.
Cognitive Interuentiom in Borderlines 213

which focuses particularly on dysfunctional mechanisms for regulating aversive

affect states and on self-damaging acts such as suicidal gestures. Linehan’s
treatment program, which includes a mixture of group therapy, individual
treatment, and psychoeducational training, tries to help borderline patients de-
velop more adequate responses for coping with affect and for solving interper-
sonal problems. In contrast to more standard cognitive therapies, Linehan (1987b)
emphasizes that “the emotional system is often the primary system and cannot
always be readily accessed or changed by modifications in the cognitive system”
(p. 152). Swenson (1989) has addressed ways Linehan’s approach may be compat-
ible and incompatible with Kernberg’s psychodynamic psychotherapy for border-
line patients. In addition to Linehan’s system, Glat (1988) has suggested that
Lazarus’ (1981) multimodal therapy may offer a more sophisticated way of
conceptualizing and treating the complexities of the borderline patient from a
cognitive-behavioral point of view.
The aim of this article is to describe integrations of cognitive-behavioral in-
terventions into the psychoanalytic psychotherapy of borderline patients. Al-
though the focus will be primarily on patients who meet DSM-III-R criteria for
BPD, many of the issues will be applicable to a broader spectrum of patients with
severe personality disorders who meet Kernberg’s (1975, 1984) criteria for border-
line personality organization. While psychotherapy integrations have been occur-
ring with increasing frequency in the past decade with other patient populations
(Goldfried et al., 1989), it is crucial that such integrations occur in the context of
integrated theories of personality, psychopathology, and technique, lest clinicians
practice a “willy-nilly” eclecticism that leads only to incoherent treatments. The
article will thus first suggest a framework for conceptualizing borderline pathol-
ogy which points to specific avenues for integration of cognitive-behavioral
techniques.

AFFECTIVE DYSREGULATION AND PATHOLOGY OF OBJECT RELATIONS IN

BORDERLINE PERSONALITY DISORDER

Two features define much of the pathology of the borderline syndrome% affective
dysregulation and pathological object relations. Identification of these two aspects
of borderline pathology, which are typically intertwined, as pathognomonic or as
generative of much of the diverse symptomatology of borderline disorders is not
new. Millon (1981), Linehan (1987a), and Klein (1977) have focused on affective
dysregulation from social learning and biological perspectives, whereas psychoan-
alytic theories (e.g., Buie & Adler, 1982; Kernberg, 1975; Masterson, 1976) tend to
focus more on the object-relational pathology of these patients. Sheehy, Gold-
smith, & Charles (1980) found, using a symptom checklist, that pathology of
interpersonal functioning, affect modulation, and impulsivity most clearly identify
borderline patients, and Clarkin (1983) identified the same three symptoms as the

4These features best define the core of “pure” borderline pathology, but they explain less of
the pathology of mixed personality disorders. Borderline-schizoid or borderline-schizotypal
patients, for example, have more prominent cognitive disturbance and less prominent
affective dysregulation. Whether peculiarities in borderline cognition (see Gartner, Hurt, &
Gartner, 1989; Silk, Lohr, Westen, & Goodrich, 1989; Zanarini, Gunderson & Franken-
berg, 1990) will constitute a third defining feature of BPD independent of the first two
remains an open question.
214 D. Westen

most frequent symptoms in borderlines. (As will be argued below, impulsivity can
be understood using a model of affect regulation and motivation.)

Affective ~y~reg~latioff in borderlines

Borderline patients manifest an extreme difficulty in regulating their emotions.

When they are upset, the world is black; when they are finding a relationship
gratifying or are craving dependence, they see the person as a savior or magical
“protector.” Perry and Klerman (1980) found affective instability to be the clearest
of several features in discriminating borderline from other psychiatric disorders.
As Masterson (1976) aptly describes the affective life of the person with BPD, “we
might say that the excursions of the pendulum in the psychological world of the
borderline patient cover an infinitely wider amplitude than in the world of the
neurotic” (p. 1’7). Ekstein and Wallerstein (1954), in describing borderline chil-
dren, proposed that a normal aspect of ego structure operates much like a
thermostat in maintaining “ego states” within a relatively consistent range. Bor-
derline patients, in contrast, appear to have a “broken thermostat,” leading to
affect states, defenses, and impulsive behaviors which spiral out of control.
Anxiety, rage, and depression centering on abandonment figure prominently in
psychoanalytic accounts of BPD (Kernberg, 1975; Masterson, 1976). As one
borderline research subject described it, “Feelings sometimes feel like they’re a
person and invade me. They push the real me off into the corner.” In recent
research interviews with patients previously reliably diagnosed as borderline and
no longer diagnosed as such years later, colleagues and I repeatedly heard
patients’ explanations of their personality change such as, ‘“I just realized I was
overreacting to everything, and I had to stop it” {see Silk, Lohr, Ogata, & Westen,
1990).
Drawing upon a number of sources (particularly Bowlby, 1969), Westen (1985)
has developed a model of affect and affect-regulation that integrates aspects of
psychoanalytic and cognitive-behavioral theories of motivation and self-regulation.
In broadest outline, affect is seen as a mechanism for the selection of behavioral,
defensive, and coping responses. A mental or behavioral process that is cognitively
associated with minimization of negative or maximization of positive affect will be
reinforced (made likely to be retained or used again) by its affective consequences.
Affect may arise in a number of ways, including, as in control theory (and Freud’s
theory of signal anxiety; see Menninger, Mayman, 8c Pruyser, 1963) as a feedback
signal to alert the person of discrepancies between emotionally invested goals and
cognized reality. This may occur consciously or unconsciously, as may selection of
responses to alter or maintain the signal (for experimental and clinical evidence,
see Westen, 1985, Chapter 2). For example, a borderline patient may feel intensely
ashamed, guilty, and afraid after verbally attacking a significant other because this
action violates a number of emotionally invested goals, including ideal-self
standards and wishes to maintain the relationship. This emotional state could
motivate any number of responses previously associated with regulation of nega-
tive affect, such as wrist-slashing (a behavior that can be used for numbing or
penance), splitting and projection (seeing the person as evil, worthless, and
provocative of the abuse-an unconscious defense), or problem-solving ways to
resolve the situation {a coping strategy). Obviously one aim of treatment with
borderline (and other) patients is to alter dysfunctional hierarchies of response
activation.
Cognitive Interventions in Borderlines 215

If one examines DSM-III and DSM-III-R criteria for BPD, one finds that many
of the criteria can be explained in whole or in part in terms of affective
dysregulation. Impulsivity or unpredictability in self-damaging ways, in large mea-
sure, reflects a deficit in affect regulation, which leads the patient to act behavior-
ally instead of mentally, to fail to delay an impulse long enough for reflection, and
to respond with behaviors as extreme as the affects motivating them. While, as will
be argued below, the pattern of unstable and intense interpersonal relationships
characteristic of borderlines is best conceived as a second core aspect of borderline
pathology, it is nonetheless intertwined with problems of affect regulation. One of
the major ways borderline patients regulate their dysphoria is by latching onto an
attachment figure for the purpose of self-soothing; they also have difficulty
staying in relationships when they get upset or do not feel gratified. Inappropriate,
intense anger or lack of control of anger is a direct example of affective dysregula-
tion, as is affective instability. While identity disturbance cannot be understood
primarily with reference to pathological affect regulation, it is related in a number
of ways. For instance, borderline patients often seem to jump from one intense
affective experience to another, precluding any subjective experience of continuity
of their own thoughts and feelings over time; indeed, because of their impulsivity
and often contradictory behavior, any observer of their behavior, borderline or
otherwise, may have difficulty piecing together a cohesive image of them. Zntoler-
ante of being alone, while best conceptualized as an aspect of borderline object-
relational pathology, also reflects the borderline patient’s deficiency in the ability
to self-soothe. Research on attachment in developmental psychology is coming
increasingly to stress the relation between affect regulation, self-soothing, and
patterns of interpersonal functioning (Kobak & Sceery, 1988). The physically
self-damaging acts and suicidal gestures characteristic of borderline patients in part
reflect their impulsivity, their experience of overwhelming depression, and their
inability to find more appropriate, tempered ways of regulating aversive affective
states. Such patients frequently explain that they cut themselves as a way of
numbing, relieving anxiety, or discharging anger. Linehan (1987a) emphasizes
such an affect-regulatory aspect of borderline cutting and “parasuicides.” The
borderline patient’s chronic feelings of emptiness and boredom may largely reflect a
failure to internalize self-soothing functions and to invest emotionally over time in
relationships, goals, and ideals. The chronic boredom of a subset of BPDs may also
reflect a residual neurologically based deficit in attention regulation (see Andro-
nulis et al., 1980), although this is speculative given the paucity of clear and
consistent findings on neurological dysfunction in BPD (see Cornelius et al., 1989).
Finally, DSM-III-R adds interpersonal exploitativeness to the borderline criteria.
While, again, this is an aspect of borderline object-relational pathology, it also
reflects an inability to invest emotionally in others in mature ways and to weigh the
needs of others when one’s own needs or affects are pressing.j

51n examining the contributions of affect-regulatory deficits in DSM-III and DSM-III-R

borderline criteria, I do not mean to suggest that all borderline symptomatology can be
reduced to this function. An affect-dysregulation hypothesis has recently been offered for
mood disorders (Siever & Davis, 1985), and both object-relational pathology and a more
general pathology of the regulation and experience of affective states (beyond depression
or even anxiety; see Benjamin et al., 1989) distinguish BPDs from patients with major
depressive disorders who are not comorbid for BPD or related personality disorders (such
as narcissistic and histrionic).
216 D. Westen

Object-Relational Pathology in Borderlines

Pathology of object relations is a second core feature of borderline psychopathol-

ogy which has been described by Kernberg, Masterson, and others. The concept of
“object relations,” however, is actually a shorthand for a constellation of cognitive
and affective processes mediating interpersonal functioning that have interdepen-
dent but distinct developmental trajectories (Westen, 1989, in press). Delineation
of specific cognitive and affective processes underlying pathological interpersonal
functioning in borderlines suggests avenues for integration of more cognitively
oriented interventions. (For empirical evidence of these processes in borderline
patients, see Westen, 1991; Westen, Lohr, Silk, Gold, & Kerber, 1990; Westen,
Ludolph, Lerner, Ruffins, & Wiss, 1990.)

Affective-~o~ivatio~af Aspects of ~order~iffe Object Re~afions

Four affective-motivational factors characterize borderline object relations. The

first is a pathology of emotional investment in relationships. Borderline patients
have unstable attachments. While at times they overinvest in someone who has
quickly become assimilated to an unrealistic schema (such as a “good mother” or
“benevolent other” of some sort), at other times they cannot maintain a consistent
attitude toward someone whom they love who is currently not being gratifying.
Second, borderline patients are aggressive (Kernberg, 1975). Whether this stems
from a history of neglect, symbiotic parental overinvolvement (see Soloff &
Millward, 1983), physical or sexual abuse (Herman, Perry, & van der Kolk, 1989;
Ogata et al., 1990; Westen, Ludolph, Misle, Ruffins, & Block, 1990), constitutional
hyperactivity or dyscontrol (Andronulis et al. 1980), or some interaction of these,
such patients’ interpersonal relations are plagued by aggressive outbursts and
sadism. Third, and related, is a characterologic masochism that typically alternates
with sadism. In many cases, this may reflect a compulsion to repeat earlier abuse,
while in other cases, the wish to be hurt reflects a developmentally primitive
mechanism to allay poorly modulated guilt. Finally, borderline patients have a
particular sensitivity to loss, rejection, and abandonment (Masterson, 1976).
Research by Perry and Cooper (1986) has begun to document interpersonal
concerns that distinguish borderlines from other patient groups.

lack of ~~~eg~atio~ of Re~resen~a~io~s of People

Borderline self- and object-representations {or, from a cognitive-behavioral per-

spective, self-schemas and person-schemas) tend to be poorly integrated (Kern-
berg, 1984; Kernberg, Selzer, Koenigsberg, Carr, & Appelbaum, 1989) and highly
variable in their complexity and differentiation. These patients have tremendous
difficulty bringing to mind representations of self and others that do not fit with
their current affect state or motivation. At times this is done defensively, as when
an adolescent patient would “forget” about positive aspects of his mother when he
would want to steal things from her. The same patient, while in the midst of a
powerful negative transference following my setting boundaries on his “emer-
gency” calls, felt so angry and rejected that he could not tolerate looking me in the
face. Instead of leaving the session, which he knew he needed, he simply turned
his chair around and began working with me as if I were the “old” Dr. Westen who
had not so badly hurt him. He seems in this case to have been able to segregate his
Cognitiue Interventions in Borderlines 217

representations of me so thoroughly by affect-tone that by turning his chair

around he could activate the more benign representation in the service of getting
help and ultimately healing the split. As will be elaborated, at other times these
patients appear to lack the capacity to maintain differentiated, integrated, and
multifaceted representations of self and others when their emotions are aroused.

Borderline Attributional Style

A third aspect of borderline object relations is what one might call a “borderline
attributional style.” In everyday life, people attribute causes to their own and
others’ behaviors, thoughts, and emotions. Research with depressed subjects has
documented the tendency of such people to have a depressive attributional or
explanatory style, in which negative events are construed as one’s own fault (an
“internal” attribution), are expected to recur (for a review, see Peterson and
Seligman, 1984). Borderline attributions tend to have the following four qualities:

1. Egocentric. Egocentrism in Piagetian theory refers to a relative lack of

differentiation of the perspectives of self and other (Piaget, 1926), or an
“embeddedness in one’s own point of view” (Looft, 1972). Borderline attri-
butions tend to be profoundly egocentric (for empirical evidence, see Exner,
1986; Westen, Lohr, Silk, et al., 1990). This egocentrism is manifest in three
ways: (a) Borderline attributions tend to be intensely personal. The self is
viewed as the causal center of the social universe. Whereas depressives tend
to make internal attributions (self as cause) of bad events and external
attributions (situation as cause) of good ones, borderline attributions are less
stable but can often be characterized as “internal-external”: “she did that
because of the way she feels about me.” (b) Because of their tendency to use
defenses such as projection and to fail to differentiate their own and others’
points of view, borderlines often attribute their own motives to others and
vice versa. (c) Borderline patients sometimes suspend attributional processes
entirely, concerned only with their own situation and feelings. For example,
one patient told his therapist, after the patient forgot about the therapist’s
vacation and appeared for his session, that if he ever came again for a session
and the therapist was not there, he would be so devastated that he would
never return. The causes of the therapist’s absence, he added, were immate-
rial because he would simply be so hurt.
2. Malevolent. Borderline patients have a tendency to attribute malevolence in
explaining events. Knight (1953) pointed to this tendency to place blame for
misfortune on external, malevolent forces. Kernberg (1975) views interpre-
tation of negative transference as critical in psychotherapy with borderlines,
in part because of their pervasive expectation that the therapist will be
malevolent.
3. Inaccurate. Borderline patients’ attributions are frequently inaccurate. Whereas,
at times, their attributions are distorted by motivational factors, at other times
this inaccuracy reflects peculiarities of borderline thinking (see Silk, Lohr,
Westen, & Goodrich, 1989; Zanarini, Gunderson, 8c Frankenberg, 1990) or a
social learning history in which parental actions were erratic and difficult to
explain (see Westen, Ludolph, Block, et al., 1990).
4. Affect-centered. Finally, borderline attributions tend to be affect-centered.
218 D. Westen

Borderlines’ attributional processes tend to become polarized by affect, with

attribution of “good” motives to “good” people and “bad” motives to “bad”
people. Borderline attributions also become global when affect is aroused.
Relatively minor events may become catastrophic because, for example, a
univalent representation is activated (e.g., “she left me because I’m worth-
less,” or “you refuse to treat me if I don’t pay for sessions because you’re a
greedy person who just likes to hurt people”). With borderline patients, as
opposed to nonborderline depressives, such globality of attribution probably
applies equally to positive affective arousal (“she’s doing that because she’s
just a wonderful person”) and to others as well as the self. Borderline
attributions are also highly susceptible to affective and motivational biases.
While this is true of normal attributional processes, the greater intensity and
peremptory quality of borderlines’ affects leads to a heightened need to
distort cognitions as well as simply a greater tendency for affective processes
to infiltrate and bias cognition. Borderline attributional style is unstable (and
difficult to measure) because it is so affect-centered and variable.

COGNITIVE-5EHAVlORAl TECHNIQUES TARGETING PATHOLOGICAL AFFECT

REGU~TION AND OBJECT RELATIONS

Most of what are called “supportive,” “ego-supportive,” or “ego-building” inter-

ventions in psychodynamic psychotherapy are aimed at helping patients regulate
their affects and impulses. Defense analysis with borderlines has this as one of its
major functions. Interpreting splitting and working with the patient to help her
see that she felt otherwise three days earlier, or that she saw the therapist or
someone else quite differently before a current idealization or devaluation, is a
technique aimed at altering a pathological affect-regulatory process (see Kernberg
et al., 1989). Reality-oriented interventions in intensive psychotherapy also serve
the function of limiting impulsive or self-damaging actions and helping the person
consider response contingencies which she was not considering while emotionally
aroused.
Because the psychoanalytic conception of certain “ego functions” is so similar to
the cognitive-behavioral concept of “self-regulation,” the kinds of interventions I
will describe may seem familiar to many psychodynamic clinicians. These tech-
niques may diverge from standard psychodynamic practice, however, in that they
are based on an analysis of the component cognitive and affective processes
believed to generate borderline symptoms -rather than on more general formu-
lations about “ego functions” or “object relations”-and hence are targeted at
specific processes. Explicit, theory-guided, consistent use of such techniques does
not necessarily follow from a conceptualization of borderline disorders as an ar-
rest at the rapprochement subphase of separation-individuation (Mahler, Pine, &
Bergman, 1975), an arrest at the developmental period in which splitting is the
primary mechanism of defense (Kernberg, 1975), a failure in the formation of a
cohesive self through transmuting internalization or optimal frustration (Buie &
Adler, 1982; Kohut, 1977), or a battleground for conflagrations among competing
object-relational part-units (Masterson, 1976). The following techniques described
represent examples of such interventions, and are clearly not intended to serve as
an exhaustive list.
Cognitive Interventions in Borderlines 219

Cognjtjve-Rebavjor~~ Strategies for Self-Reguiatjon

Many cognitive-behavioral techniques are well-suited to borderline patients be-

cause of their explicit focus on self-regulation (see Kanfer & Karoly, 1972; Karoly
& Kanfer, 1982). Meichenbaum’s (1977) approach to teaching impulse control to
impulsive children, for example, focuses on the progressive internalization and
routinization of self-instructions to stop, look, listen, delay, plan, and implement
an appropriate strategy. Various sequenced directions of this sort are at first
provided concretely by the therapist, but this function is gradually assumed by the
patient, who at first instructs himself verbally and ultimately goes through these
steps unconsciously.
Certain standard cognitive techniques, such as those directed against “catastro-
phizing,” or “black-and-~~hite” or “either-or” thinking (Beck, 1976), are particu-
larly appropriate for treating borderline affect regulation. The therapist points
out the way these cognitive patterns are activated under certain circumstances and
helps the patient reality-test them when they emerge. Borderline patients are not
easily “argued out of’ their distortions, and are seldom able when their emotions
are aroused to engage in the kind of reality-oriented “collaborative empiricism”
(Beck, 1985) of cognitive therapy. Consequently, the therapist needs to respond
empathetically to the patient’s intense affect state while simultaneously addressing
distortions that may both reflect and intensify it (e.g., “I know it feelslike that, but
is that really likely?“).

Red Hags and 8/ack Dots

One patient (a borderline adolescent) came to develop, in therapy, a list of “red

flags” which signal to him that a pathological process is in progress. We began the
list with a series of words which, if he heard himself utter them, should serve as a
red flag. The list included “bitch,” “asshole,” “perfect,” “total _” and other
words that signaled all-or-nothing thinking and spiraling affects. Gradually, he
came to add other psychological processes and behaviors to his list of red flags. For
example, when he felt himself wanting to use cocaine, he would recognize that he
was running from an intense affect or expressing a dynamic behaviorally, partic-
ularly since his father’s death had been cocaine related. Whereas, initially, over the
course of many sessions, I had to point out red flags repeatedly, he eventually
began noticing them in his daily life as well as in sessions.
While identification of red flags was a necessary first step, the patient somewhat
tauntingly asked me the first time we began using this terminology, “So then what?
What am I supposed to do with it.2” I wondered with him what he and I do
together when his affects begin to spiral out of control in sessions, to which he
replied, ‘*Be analytical.” This, then, became the second step after identifying a red
flag, namely, inhibiting action long enough to examine his associations and the
meanings, fears, wishes, or conflicts generating the emotion. Doing so slowed him
down, which is critical for borderline patients, who have difficulty inhibiting action
and exercising self-observation when affectively aroused. It also fostered a com-
forting identification with the therapist which was itself soothing and helped him
respond appropriately to whatever concern caused the emotion rather than
responding globally and diffusely to what felt like a blinding affective signal.
It should be noted that this technique, which relies on insight in the psychoan-
220 D. Westen

alytic sense, was much more concretely worked out with the patient than psy-
chodynamic clinicians are usually comfortable doing. For example, we spoke
directly in terms of a two-step process and gave labels to each of the steps.
Sometimes when he was not in a session but began to feel his affects spiraling out
of control, he was encouraged to write my “part” in the therapeutic dialogue, or
ask himself questions as I would. When in sessions he would start to spiral and
helplessly flail around waiting for me to find the magic to calm him, I was initially
very explicit in helping him begin to assume some of my functions by asking what
I would say or ask right now. While borderline patients are very sensitive to
condescension or to “techniques” that do not feel like part of a genuine relation-
ship, concreteness is of the essence when their affects are poorly modulated
because their capacity to think abstractly and to sequence thoughts or actions
becomes grossly impaired. Kernberg and his colleagues (1989) similarly advocate
labeling role relationships that repeatedly emerge in the treatment. Ideally, the
need for concreteness, as in speaking of red flags or using labels for various
processes or “states of mind” (Horowitz, 1987), should gradually and spontane-
ously subside, as the patient internalizes the process and routinizes the procedure
of spotting and responding to dysregulation (although when severely stressed, the
patient may once again require concrete interventions).
Another patient with mixed borderline-narcissistic pathology experienced imag-
ined interpersonal slights with such intensity that “the world” would become an
undifferentiated evil mass that included stupid, malevolent, faceless people out to
make him suffer. At such times he was very difficult to reach because the therapist,
too, was part of this undifferentiated mass. Eventually, I employed the metaphor
with him that when his depression, rage, or humiliation became aroused, the
world would collapse into one big “black dot, ” and that he needed to let me out of
the dot momentarily so we could discuss it. With repeated use of this metaphor he
began to have some recognition of this process as it would occur in the therapy
hours, and would spontaneously offer, “I’m black-dotting again.” “Red flags” and
“black dots” are, of course, simply examples of the way structured interventions
may emerge from metaphors developed in treatment hours.

Affective-motivational aspects of object relations pathology may be addressed

through careful use of modeling, as when a patient tries to “cut a deal” regarding
some slightly shady handling of third-party payment. In such cases, one can
effectively both examine the meaning of the patient’s attempt to draw the therapist
into collusion with a frauduient action and also demonstrate to the patient a way
of handling ethical issues and inhibiting need-gratification in a way that may be
quite foreign to the patient’s experience. Although such interventions are often
understood in psychoanalytic psychotherapy as “limit setting” (i.e., as treatment
parameters), they can also be very valuable in modeling ways of handling impulses
or emotions. Modeling affect-regulatory processes can also be useful. For example,
the adolescent described above had a penchant for pushing all the buttons on the
elevator on his way up to my office while he was an inpatient. One day, when he
did this, he evidently “pushed my buttons” as well, since annoyance apparently
flashed across my face. He, of course, was quite attuned to any signs that would
Cognitive Interventions in Borderlines 221

indicate that he had achieved the desired result, which he also feared. In the
process of exploring the meanings of his testing me this way, I disclosed that I had,
indeed, been annoyed, but that I had then begun thinking about several possible
explanations of his behavior. He was astonished both that a person could pause
and self-reflect when getting angry and that one could still think clearly enough to
generate dynamic hypotheses in such circumstances. Similarly, whether one thinks
of this in terms of modeling or identi~cation, the patient’s internalization of the
therapist’s relative calm and analytic attitude in the face of his or her escalation can
be an important source of therapeutic change.

Limited Self-Disclosure

While one must be extremely careful to examine both countertransference motives

and transference meanings whenever one self-discloses to a borderline patient,
judicious use of self-disclosure can at times be helpful and, at other times, in-
dispensible. Distorted attributions, often of a malevolent nature, are pathogno-
manic of borderline pathology and of personality disorders more generally, and
these invariably arise in the treatment situation. The therapist’s disclosure of his or
her actual motives alongside exploration of the meanings and causes of the
patient’s attributional distortions can be crucial in helping the patient reality-test
his or her inferences; it can also sometimes help prevent premature termination
resulting from malevolent attributions.
Borderline personalities are often products of chaotic homes in which parental
figures behaved inexplicably and erratically (see Ludolph et al., 1990). Therapists
whose methods remain mysterious or ambiguous may be playing into pathological
family dynamics as well as childhood fantasies. Particularly with these patients,
whose boundaries were often violated psychologically or sexually as children,
initial transference interpretations should usually be prefaced with some kind of
explanation of why the therapist is focusing on this relationship at all (such as,
“What you feel in here with me can sometimes be very similar to what you feel with
other people in your life, so it can often be helpful to look at what’s happening in
this relationship,” or “What a person feared or hoped for in his relationships with
his parents as a child often seems to come up again in his relationships with other
people later on. I wonder if the way you seem to have such a hard time trusting me
has anything to do with the way you talked about feeling that you couldn’t rely on
anyone when you were a kid.“). Borderline patients usually will not, however,
maintain appropriate boundaries in the therapeutic relationship, so one must be
very circumspect with self-disclosure.

Humor

A well-timed and .judicious use of humor can also be valuable in working with
borderline patients, not only in helping build an alliance, but in counteracting
runaway affect. The role of the therapist in this instance is something like the role
of government in Keynesian economics; i.e., when normal equilibration processes
cannot prevent a downward and self-reinforcing spiral, an exogenous factor must
intervene to prevent depression. Humor can also be helpful at times in address-
ing manipulations, because it signals to the patient that the therapist knows what
222 D. Westen

the patient is trying to do and will not go along with it, but neither will he
be destroyed by it nor rejecting in response to it.

Enhancing Perspective-Taking

Much of the work therapists do with borderline patients centers on enhancing

capacity for perspective-taking, particularly under conditions of affective arousal.
The tendency of these patients to think egocentrically requires specific therapeutic
focus, as when the therapist questions the patient’s view that “everyone” in his life
is out to hurt her, or helps the patient see why her employer was distressed when
she persistently called in sick with ten minutes’ notice. Such interventions are
particularly useful in dealing with transference issues because these issues are
affective and live, both partners in the experience have firsthand knowledge of the
interactions involved, and the therapist can directly confront and, at times,
contradict the patient’s attributions.

Examining Cognitive Structures, Patterns, and Processes

Interventions aimed at increasing the complexity, integration, and differentiation

of representations of people and at altering pervasive, dysfunctional attributional
patterns are typically cognitive in focus but are a standard part of psychoanalytic
psychotherapy with borderline patients. Kernberg and his colleagues (1989) have
masterfully described the way one addresses simplistic and poorly integrated
representations. At times labeling a problematic cognitive pattern directly can be
helpful, with borderlines as well as with other personality disorders. One patient
had the pervasive fantasy that she was always about to hit the last straw with people
which would convince them that she was stupid, worthless, and should therefore
be abandoned. Not surprisingly, this served as a tremendous resistance to treat-
ment, since she censored every word for fear that the next utterance would lead to
this unfortunate state of affairs. This “last straw” fantasy rested upon both a
simplistic, overly global self-representation (“I am worthless”), a masochistically
distorted attribution (“if this person hasn’t abandoned me yet, it is because he
hasn’t noticed my awfulness yet”), and an egocentric attribution (“everyone sees
me the same way I do”). As in a traditional psychoanalytic treatment, the
masochistic dynamic was examined in terms of the conflicts underlying it. On the
more cognitive side, we came to label the egocentric pattern of attribution “Ms.
C’s First Law,” namely that people see things exactly as she does, and tested the
reality of her attributions on numerous occasions when she or I became aware that
she was applying this “law.” The lack of complexity of her self-representation,
particularly when she was depressed, was examined both cognitively (looking at
how this simplistic schema affects her life) and dynamically (looking at her
motivations at particular times for evoking it). On one occasion, when she was
most distraught and terrified that her next “stupid” words would be the last straw,
I graphically illustrated a component of my own schema of her, which included
numerous hierarchically organized subcomponents, and noted the way a single
comment at any point would be assimilated into this structure. I pointed out a
critical difference between her cognitive structure and mine, that mine was not so
brittle or one-dimensional that it would collapse in the face of a few utterances on
her part. This intervention allowed her to begin discussing the pathologically
harsh moral standards that paralleled her brittle self-representation, as well as her
fear that some of her comments were not only stupid but aggressive, and hence
Cognitive Interuentions in Borderlines 223

back into a more analytic examination of a number of very rigid impulse-defense

constellations. While the use of a graphic illustration in psychoanalytic therapy is
certainly unconventional (although this patient was an adolescent), it can be
effective when affect is intense and thinking has regressed to a more concrete
level. Confrontation of other attributional patterns is standard fare in the psycho-
analytic treatment of borderline patients, although not always conceptualized as
such. For example, altering the pervasive expectations of victimization is perhaps
the most critical aspect of transference analysis advocated by Kernberg with
borderline patients.

Integration of Multiple Technical Strategies

Addressing problematic attributional processes, representations of self and others,
and dynamic conflicts often occurs in tandem. One patient, who had tremendous
difficulty with self-esteem, would, when frustrated, apply to herself overly simplis-
tic and undifferentiated trait labels such as “stupid.” She was unable in such
moments to remember ways in which she was not all-dumb or to generate
alternative hypotheses for failures. On one occasion she had recently returned to
college after taking a high school equivalence examination, and decided after
doing poorly on a test that she was stupid and would never complete the term. She
became terribly distraught, viewing this failure as one more instance of her
worthlessness. Her wish to do well, and her profound disappointment at her
grade, led not only to activation of a global and univalent self-schema but also to
an abridgement of the attribution process (i.e., a failure to consider other
explanations). What was most useful for her in the session, and what allowed her
to calm down and begin to cope more effectively (e.g., by planning to go to the
library that night instead of picking up a man at a bar, as she had intended), was
my pointing out the activation of a global and univalent self-representation, and
wondering about her failure to consider the impact on her grade of her never
having completed high school, which would have tempered her attribution of
stupidity as the sole explanation for failure. (I was comfortable in intervening in
this way, which could be interpreted as simply bolstering a rationalization, because
I had her WAIS-R protocol in my file drawer.)
These kinds of interventions frequently lead back to more dynamic interven-
tions. For example, this patient, having calmed considerably in the session, did not,
in fact, go to the library as planned and rapidly began fretting about failure on the
next test. Examination of her failure to follow through revealed her fear that she
could not meet my expectations of success, just as she felt she could not meet her
father’s as a child, and her consequent attempt to bring failure under her control
by doing something to bring it about. She promptly acted herself out of school and
then confronted me with precisely the situation with which she had confronted her
father so many times, namely that she had raised his hopes and then dashed them.
In my stance toward school, I had unthinkingly moved into the role of an advocate
and thus clearly diverged from neutrality, and in this case, I paid for it. Through
this experience, however, the patient gained insight into the way she enlists
“accomplices” (Wachtel, 1987) who recapitulate with her many painful experiences
from her past.
Use of cognitive-behavioral techniques in the context of intensive psychotherapy
neither interferes with transference-oriented work nor does it irrevocably “con-
taminate the transference.” Treatment of the adolescent boy described here in
224 D. Westen

various examples, for instance, centered not only on examination of his alternately
idealizing, devaluing, manipulating, and feeling victimized by the therapist, but
also on a prominent homosexual transference. Therapeutically, one can work with
a patient’s experience or distortions of the implications of various interventions
on the part of the therapist, just as one works with patients’ interpretations and
distortions of a relatively neutral stance. For example, careful analytic work with a
patient who experiences a technical decision as manipulative may uncover a
projection of his own wishes to manipulate, an instance of devaluation or
expectation of victimization, or a tendency to treat an overdetermined action as if
it were the product of a single (usually hostile) motivation. Further, transitory
assumption of a slightly different role by the therapist may allow access to more of
the patient’s transferential paradigms. Wachtel (1987) has cogently discussed
transference implications of behavioral interventions in psychoanalytic psycho-
therapy in an essay on the limits of therapeutic neutrality entitled, “You Can’t Go
Far in Neutral.” Therapists who work with borderline patients are always caught
between the Scylla of recapitulating neglectful or affectively ingenuine parenting
through what the therapist experiences as neutrality, and the Charybdis of recapit-
ulating poorly bounded, overinvolved parenting motivated by countertransference
feelings. Unfortunately, the course must be charted anew for each patient, with a
very imperfect compass.

THE LIMITS OF PSYCHOANAL~IC AND COGNITIVE-BEHAVIORAL

INTERVENTIONS WITH BORDERLINE PATIENTS

As this article has suggested, consideration of cognitive-behavioral interventions

may suggest avenues of intervention not typically emphasized in the psychody-
namic psychotherapy of borderline patients. The need for cognitive-behavioral
supplementation of this sort stems from some of the limits of the psychodynamic
paradigm.

Limitations of Psychoanalytic Approaches

First, while psychoanalysis proposes that conscious uncovering of unconsciously

evoked patterns (such as defenses, conflicts, and relationship paradigms) is critical
to therapeutic change, it lacks a comprehensive theory of the way conscious
decisions in the face of insight become transformed into adaptive automatic or
unconscious patterns. Whereas in psychoanalysis, automaticity is often associated
with pathology, in cognitive psychology automaticity is viewed as a prime charac-
teristic of skilled functioning which distinguishes experts from novices (see
Anderson, 1985). In the case of borderline patients, mere conscious recognition of
a pattern, even in the face of strong affect (or perhaps, especially in the face of
strong affect, in contradistinction to neurotics), does not typically lead to change
in fundamental patterns of functioning (such as affect regulation) without consid-
erable practice, trial and error, and conscious structuring of mental processes.
Psychoanalytic theory and technique focus primarily on one important class of
affect-regulatory mechanisms-unconscious defenses-without adequately address-
ing conscious techniques for self-regulation other than “insight.”
Second, psychoanalytic theory and practice tend not to focus adequately enough
on implementation of change, which does not typically follow automatically from
insight, even in neurotics. Insight into the causes of an irrational fear, for example,
Cognitive Interuentims in, Borderlines 225

may alter the balance among anxiety, wishes to approach the feared object,
expectations of efficacy in overcoming the fear, and expectancies about the
dangers involved in doing so. However, insight alone often will not provide
enough impetus to approach the object and, hence, to begin disconfirming the
anxious expectancies or to inhibit automatic elicitation of an affective reaction
associatively connected to a cognitive representation of the phobic stimulus (see
Wachtel, 1987). Similarly, insight into the neurotic motivations for depressive
self-denigration may not be entirely effective in inducing the person regularly to
access positive self-representations because thinking positively about the self is
linked to aversive affects and cognitive-affective schemas that are activated auto-
matically (such as fears that if one does not blame oneself one must attribute blame
instead to an abusive parent, who has been defensively shielded from responsibil-
ity or idealized). In such cases, thinking positive thoughts about the self is
tantamount to a phobic stimulus that must be “desensitized.” Implementation of
insight-driven change must be a particular target of therapeutic work with
borderline patients, whose counter-motivations and difficulty implementing nny-
thing can often derail a treatment.
Third, the psychoanalytic preference for motiva~onal explanations underplays
the role of cognition and social learning in the etiology and maintenance of
psychopathology. Kernberg, for example, views splitting as a defense against the
anxiety evoked by forming more complex representatives (Kernberg et al., 1989),
as when a patient refuses to see any imperfection in his therapist to avoid
intolerable feelings of disappointment and rage. While splitting is often motivated
in this way, this is only one explanation of what is probably a multidetermined
phenomenon. Splitting could result from deficits in affect regulation without any
defensive function if strong affects recruit univalent or mood-congruent rep-
resentations. Research suggests that mood-states broadly influence cognition,
including encoding and retrieval of schematic information (Blaney, 1986; Glligan
& Bower, 1984; Isen, 1984). An inability to regulate rage or sadness, for example,
could produce a tendency to make malevolent attributions or to activate powerful
and global negative self-representations and could inhibit recognition of alterna-
tive or more complex explanations. Splitting can also reflect social learning. The
patient described earlier wit.h the “last straw” fantasy had a severely borderline
mother (whom I knew to be so by direct observation} whose representations were
so split and transitory that the patient literally never knew whether she would be
facing the “totally loving mother” or the “totally attacking mother,” or whether she
(the child) would do something that would make her totally had in her mother’s
eyes - the “last straw.” Parental or familial projections onto the child and other
family members may similarly lead to internalization of rigid roles and split views
of the self and others {such as “the evil child”) {see Shapiro et al., 1975).
Fourth, the confusing place of affect in psychoanalytic theory (see Rapaport,
1953) has led to a relative failure to focus theoretically on a core aspect of
borderline pathology, namely the experience and regulation of affect. Theoretical
and clinical attention have instead focused on drive considerations and object-
relational formulations which, though useful, draw attention away from what
should be an equally important and discrete focus of therapeutic attention.
Finally, the holism endemic to psychoanalysis, which views character as a system
of interdependent processes and understands symptoms as particular manifesta-
tions of more general conflicts or personality dynamics, focuses the clinjcian’s
attention away from specific processes (e.g., particular aspects of borderline
226 D. Westen

attributions) and onto more general phenomena (e.g., “object-relational pathol-

ogy* ” “separation issues”). In Kernberg’s system, all borderline symptoms are
ultimately derived from the patient’s defensive efforts to keep good and bad
representations separate (see Kernberg et al., 1989). Holism, in this respect, is one
of the strengths as well as weaknesses of psychoanalytic theory, just as the atomism
and attention to microprocesses inherent in cognitive-behavioral theory and
technique have both costs and benefits.

limitations of Cognitive-Behavioral Approaches

While cognitive-behavioral approaches can be usefully integrated into psychoana-

lytic psychotherapy, as an alternative to intensive psychotherapy they have limita-
tions which are particularly salient in work with personality-disordered patients,
especially borderlines. First, cognitive-behavioral approaches underestimate the
complexities of motives and countermotives, For example, at a recent conference,
a prominent cognitive-behavioral clinician was asked to comment on a case of an
adolescent diabetic from a broken home who kept throwing himself into ketoaci-
dosis. After recommending education about diabetic control and being apprised
that this boy knew more about diabetes than his parents or his therapist and yet
still kept hiding candy bars, the speaker suggested that perhaps peer pressure
motivated the boy to want to hide his diabetes in order to be like his peers. This
level of understanding of motivation is not far advanced of American popular
culture; it ignores the complex motives of a boy who had been physically abused,
had abused himself on numerous occasions and developed self-abusive motiva-
tions, and seemed to be tempting fate to prove his immortality in the face of a
serious illness. The cognitive-behavioral understanding of motivation, with its
pragmatic, rationalistic bent inherited from the Enlightenment, British empiricist
philosophy, and twentieth century American culture, is particularly inadequate in
dealing with borderline patients. Anyone who has used operant techniques with
inpatient borderlines is aware of the paradoxes of ‘~rei~~forcement” in patients with
strong masochistic trends who are often “reinforced” by punishment.
Second, and related, cognitive-behavioral approaches considerably underesti-
mate the complexities of fantasy and unconscious experience. One suicidal
borderline patient told me in the context of a research interview that she had
recently been in treatment but had not disclosed to her therapist any of the
aggressive fantasies which tormented her, and which she had disclosed to me in a
single interview. When asked why not, she said she is not a nice person, whereas
her therapist is, and this was a “nice therapy” in which one did not discuss such
things. Similarly, a patient who was evaluated for “trouble getting along with
peopie” provided indications in the first interview of a perversion. He had several
very gruesome sexual fantasies, some of which involved his prior therapist, a
cognitive therapist he had seen for over a year. He had numerous fantasies of
raping and killing her which he had never discussed with her and about which she
had never asked. Instead, I received a treatment summary from her that glowingly
described his progress in their work on his “social skills” and “workaholism.” In
treating severe personality disorders, psychotherapy that stays at the level of skills
training and role-playing to overcome problems such as “relating to people” and
“workaholism” is particularly problematic because these patients are often terrified
of their fantasies and are certain that the therapist is equally afraid of dealing with
Cognitive Interventions in Borderlines 227

them. The focus of cognitive approaches is primarily on conscious or semicon-

scious habits of thought, with minimal attention to deeper levels of meaning and
experience and to possible motivations for these habits. Cognitive theories also lack
a concept of defense, addressing instead only conscious coping mechanisms, which
are only one form of affect-regulatory mechanism.
Third, cognitive-behavioral theories lack a concept of levels of object relations
and ego functioning which permit a distinction between neurotic symptoms and
severe character pathology. In cognitive-behavioral literature (at least until very
recently), depression is depression, phobias are phobias, and no further diagnostic
distinctions are necessary for treatment. Sophisticated cognitive-behavioral clini-
cians recognize these distinctions and adapt their treatment strategies accordingly,
but they do so without the aid of theory or of any conceptualization of character
structure. Less experienced clinicians operating in this theoretical no-man’s_land
with character disorders are prone to making gross technical errors, as in the case
of a borderline patient referred to me after a suicide attempt who had been
treated solely with relaxation training, or a narcissistic patient whose aggressive
outbursts and difficulty attaching made him friendless and jobless who was treated
with cognitive therapy aimed at helping him see that he is not as unlovable and
unemployable as he accurately recognized himself to be.
Fourth, although attention to specific pathological processes is one of the
strengths of the cognitive-behavioral approach, it is also a weakness when one
begins to consider interactions of various cognitive and affective processes that may
render treatment of any one process alone highly problematic. Targeting for
treatment the presumed perspective-taking deficits of delinquents or antisocial
personalities may be of little value if one does not simultaneously target the
pathology of inaestment in relationships that would make such individuals cure about
the perspective of others. Cognitive “deficits” of this sort may be only one part of
a broader pathology of attachment involving a system of interacting cognitive,
affective, and motivational processes.
Finally, although cognitive concepts could prove useful in understanding cer-
tain transferential processes (Goldfried, 1989; Singer, 1985; Wachtel, 1981;
Westen, 1988), cognitive-behavioral approaches fail to address the complexities of
the interpersonal process between patient and therapist. For borderline patients,
much of whose pathology is interpersonal, this imposes a severe limitation. This
can be seen in the failure of the therapist in a previously cited example to examine
the patient’s sadistic fantasies toward her, which recur in other relationships, or
the failure of the “nice therapist” to examine the patient’s object-relational
paradigm of hating herself, protecting a significant other from her destructive
wishes, and consequently remaining isolated from people with whom she longs to
be intimate but simultaneously keeps at arm’s length.
The field of psychotherapy has long been polarized between advocates of
psychodynamic and cognitive-behavioral approaches. Perhaps it would be thera-
peutic for both patients and the profession for some healing of this split to take
place in the treatment of patients for whom splitting and polarization are a
prominent aspect of their pathology.

Acknowledgements - The author would like to thank Christopher Peterson, Kenneth Silk,
Paul Wachtel, and several anonymous reviewers for their comments on a draft of this
paper.
228 D. Westen

REFERENCES

Anderson, J. R. (1985). Cognitive psychology and its implications (2nd ed.). New York: Freeman.
Andronulis, P. A., Clueck, B. C., Stroebel, C. F., Vogel, N. G., Shapiro, A. L., & Aldridge, D. M. (1980).
Organic brain dysfunction and the borderline syndrome. Psychiatric Clinics of North America, 4,47-66.
Benjamin, J., Silk, K. R., Kohr, N. E., & Westen, D. (1989). The relationship between borderline
personality disorder and anxiety disorders. American ,/oumnl of Orthopsychiatq, 59, 461-467,
Beck, A. I’. (1976). Cognitive therapy and the emotional dzsorders. New York: International Universities
Press.
Beck, A. T. (1985). Cognitive therapy. In H. I. Kaplan & B. J, Sadock (Eds.), Comprehensive textbook of
psychiatry (pp. 1432-1438), 4th ed. Baltimore: Williams and Wilkins.
Beck, A. T., Freeman, A., et al. (1990). Cognitive therapy ofppersonalzty disorders. New York: Guilford
Press.
Beitman, B. D., Goldfried, M. R., 8c Norcross, J. (1989). The movement toward integrating the
psychotherapies: An overview. Am&can Journal ofPsychiat9, 146: I38-147.
Blaney, P. H. (1986). Affect and memory: iz review. Ps~~~iolo~~al Et&&z, 99, 229-246.
Blatt, S. J., Quinlan, D. M., Chevron, E. S., McDonald, C., & Zuroff, D. C. (1982). Dependency and
self-criticism: Psychological dimensions of depression. Journal of Consultq and ~l~~~cal Ps~cholo~, 50,
113-124.
Bowlby, J. (1969). Attachment and toss: Vol. I, Attachment. New York: Basic Books.
Buie, D., & Adler, G. (1982). The definitive treatment of the borderline patient. Internationnl,Jaurnal of-
Fsychoana~yt~c Psychot~rapy, 9, 51-87.
Clarkin, J. F. (1983). Prototypic typology and the borderline personality disorder. JournaE of Ab~o~~~
Fsycholo#, 92, 263-275.
Cowdry, R. W., & Gardner, D. L. (1988). Pharmacotherapy of borderline personality disorder, Archives
of Gefteral Psychiatry, 45, 11 l-l 19.
Ekstein, R., & Wallerstein, J. (1954). Observations on the psychology of borderline and psychotic
children. Psy~oanalyt~c Study of the Child, 9, 344-369.
Exner, J. E. (1986). Some Rorschach data comparing schizophrenics with borderline and schizatypal
personality disorders. Journal of Personality Assessment, 50, 455-47 1.
Freeman, A., & Leaf, R. C. (1989). Cognitive therapy applied to personality disorders. In A. Freeman,
K. M. Simon, L. E. Beutler, & H. Arkowitz (Eds.), Comprehensive handbook of cognitive therapy, (pp.
403-433). New York: Plenum.
Gardner, D. C., & Cowdry, R. W. (1985). Alprazolam-induced dyscontrol in borderline personality
disorder. American Journal of Psychintry, 142, 98-100.
Gartner, J., Hurt, S. W., & Gartner. A. (1989). Psychological test signs of borderline personality
disorder: A review of the empirical literature. Jou.rnal of Personality Assessment, 53, 4 13-42 1.
Gilligan, S. G., & Bower, G. H. (1984). Cognitive consequences of emotional arousal. In C. E. Izard, J.
Kagan, & R. B. Zajonc (Eds.), Emotion, cognztzon, and behavior. Cambridge: Cambridge University
Press.
Glat, M. (1988). Borderline syndrome: Cognitive and behavioral treatment for adolescents. Part
II-Treatment and practice. Car&r Foundation Letter, 132, 1-3.
Goldberg, S. C,, Schultz, P. M., Resnick, R. J., Hamer, R. B., & Fried& R. 0. (1986). Borderline and
schizotypal personality disorders treated with low-dose thiothixene vs. placebo. Archives of General
Psychzatry, 43, 680-686.
Goldfried, M., & Hayes, A. (1989). Can contributions from other orientations complement behavior
therapy? The Behavior Therapist, 12, 57-60.
Gunderson, J. C. (1984). Borderline persona&y dzsorder. Washington, DC: American Psychiatric Press.
Gunderson, J. G. (1986). Pharmacotherapy for patients with borderline personality disorder. Archzves oif
General Psych~t~, 43, 698-700.
Gunderson, J. G., Rc Elliott, G. R. (1985). The interface between borderline personality disorder and
affective disorder. American Journal of Psych~t~, 142, 27?-288.
Herman, J.. Perry, J. C., & van der Kolk, B. A. (1989). Childhood trauma in borderline personality
disorder. ‘4~~~~~~ Jou~al of Psychia&~, 146, 490-495.
Horowitz, M. (1987). Stat&s of mind: Co~f~~&~at~o~.a,l anaEysk of ~~2d~vidz~al psycl~o~o~ (2nd Ed.), New York:
Plenum Press.
Hunt, H. F. (1977). Behavioral perspectives in the treatment of borderline patients. In P. Hartocollis
(Ed.), Borderline personality disorders. New York: international Universities Press.
Isen, A. (1984). Toward understanding the role of affect in cognition. In R. S. Wyer & T. K. Srull
(Eds.), Handbook of socznl cognition, Vol. 3. Hillsdale, NJ: Erlbaum.
Cognitive Interuentions in Borderlines 229

Kanfer, F. H., & Karoly, P. (1972). Self-control: A behavioristic excursion into the lion’s den. Behavior
Therapy, 3, 398-416.
Karoly, P., & Kanfer, F. H. (Eds.) (1982). Self- management and behavior change: From theory to p-a&ice. New
York: Pergamon Press.
Kernberg, 0. (1975). Borderline conditions and pathological narcissism. New York: Aronson.
Kernberg, 0. (1984). Severe personality disorders: Psychotherapeutic strategies. New Haven: Yale University
Press.
Kernberg, 0. F., Selzer, M. A., Koenigsberg, H. W., Carr, A. C., & Applebaum, A. H. (1989).
Psychodynamic psychotherapy of borderline patients. New York: Basic Books.
Klein, D. (1977). Psychopharmacological treatment and delineation of borderline disorders. In P.
Hartocollis, (Ed.), Borderline personality disorders: The concept, the syndrome, the patient (pp. 365-384).
New York: International Universities Press.
Knight, R. (1953). Borderline states. Bulletin of the Menninger Clinic, 17, l-l 2.
Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation,
and representations of self and others. Child Development, 59, 135-146.
Kohut, H. (1977). The restoration of the self. New York: International Universities Press.
Kroll, J., & Ogata, S. (1987). The relationship of borderline personality disorder to the affective
disorders. Psychiatric Developments, 2, 105-128.
Lazarus, A. A. (1981). The practice of multimodal therapy. New York: McGraw-Hill.
Lehrer, P., Schiff, L., & Kris, A. (1971). Operant conditioning in a comprehensive treatment program
for adolescents. Archives of General Psychiatry, 25, 5 15-52 1.
Liebowitz, M. R., Stone, M. H., & Turkat, I. D. (1986). Treatment of personality disorders. In A. J.
Frances & R. E. Hales (Eds.), American Psychiatric Association Annual Review, Vol. 5. Washington, DC:
American Psychiatric Press.
Linehan, M. (1987a). Dialectical behavior therapy for borderline personality disorder: Theory and
method. Bulletin of the Menninger Clinic, 51, 261-276.
Linehan, M. (1987b). Dialectical behavior therapy in groups: Treating borderline personality disorders
and suicidal behavior. In C. M. Brady (Ed.), Women in groups. New York: Springer.
Linehan, M. (1987c). Dialectical behavior therapy: A cognitive behavioral approach to parasuicide.
Journal of Personality Disorders, 1, 328-333.
Looft, W. R. (1972). Egocentrism and social interaction across the lifespan. Psychological Bulletin, 78,
73-92.
Ludolph, P., Westen, D., Misle, B., Jackson, A., Wixom, J., & Wiss, F. C. (1990). The borderline
diagnosis in adolescents: Symptoms and developmental history. American Journal of Psychiatry, 147,
470-476.
Mahler, M., Pine, F., & Bergman, A. (1975). The psychological birth of the human infant: Symbiosis and
individuation. New York: Basic Books.
Masterson, J. F. (1976). Psychotherapy of the borderline adult. New York: BrunnerlMazel.
Meichenbaum, D. (1977). Cognitive behavior modification: An integrative approach. New York: Plenum
Press.
Menninger, E., Mayman, M., & Pruyser, P. (1963). The vital balance. New York: Viking Press.
Millon, T. (1982). Disorders of personality: DSM-III, Axis II. New York: John Wiley & Sons.
Ogata, S., Silk, K. R., Goodrich, S., Lohr, N. E., Westen, D., & Hill, E. (1990). Childhood abuse and
clinical symptoms in borderline personality disorder. American Journal of Psychiatry, 147, 1008-1013.
Perry, J. C., & Cooper, S. H. (1986). A preliminary report on defenses and conflicts associated with
borderline personality disorder. Journal of American Psychoanalytic Association, 34, 863-893.
Perry, J. C., & Klerman, G. (1980). Clinical features of borderline personality disorder. American Journal
of Psychiatry, 137, 165-173.
Peterson, C., & Seligman, M. E. P. (1984). Causal attributions as a risk factor for depression: Theory
and evidence. Psychologtcal Revzew, 91, 347-374.
Piaget, J. (1951). The language and thought of the child. New York: Humanities Press (Original work
published, 1926).
Pilkonis, P., & Frank, E. (1988). Personality pathology in recurrent depression: Nature, prevalence, and
relationship to treatment response. American Journal of Psychiatry, 145, 435-441.
Rapaport, D. (1953). The psychoanalytic theory of affect. In M. Gill (Ed.), The Collected Works of David
Rapaport. (pp. 476-512). New York: Basic Books.
Shapiro, E. R., Zinner, J., Shapiro, R. L., & Berkowitz, D. A. (1975). The influence of family experience
on borderline personality development. International Review of Psycho-Analysas, 2, 3999411.
Sheehy, M., Goodsmith, I., & Charles, E. (1980). A comparative study of borderline patients in a
230 D. Westen

psychiatric outpatient clinic. American Journal of Psychzatry, 13’7, 1374-1379.

Siever, L. J., & Davis, K. L. (1985). Toward a dysregulation hypothesis of depression. American Journal
of Psych&at?, 142, 1017-1031.
Silk, K. R., Lohr, N. E., Westen, D., & Goodrich, S. (1989). Psychosis in borderline patients with
depression. Jounal of Personalzty Dzsorders, 3, 92-100.
Silk, K. R., Lohr, N. E., Ogata, S., & Westen, D. (1990). Borderline inpatients with affective disorder:
Preliminary follow-up data. Journal of Personality Dzsorders, 4, 213-224.
Singer, J. (1985). Transference and the human condition: A cognitive-affective perspective. Psychoan-
alytic Psychology, 2, 189-2 19.
Soloff, R., & Millward, J. W. (1983). Developmental histories of borderline patients. Comprehensive
Psychiatry, 24, 574-588.
Soloff, P. H., George, A., Nathan, R. S., Schulz, P. M., Ulrich, R. F., & Perel, J. M. (1986). Progress in
pharmacotherapy of borderline disorders: A double-blind study of amitriptyline, haloperidol, and
placebo. Archives of General Psychzatq, 43, 691-697.
Swenson, C. (1989). Kernberg and Linehan: Two approaches to the borderline patient. Journal of
Personality Disorders, 3, 26-35.
Wachtel, P. (1981). Transference, schema, and assimilation: The relationship of Piaget to the
psychoanalytic theory of transference. Annual of Psychoanalysis, 8, 59-76.
Wachtel, P. (1987). Znsight and change. New York: Guilford.
Waldinger, R. J. (1987). Intensive psychodynamic therapy with borderline patients: An overview.
American Journal of Psych&y, 144, 267-274.
Waldinger, R. J., & Gunderson, J. G. (1984). Completed psychotherapies with borderline patients.
Amerzcan Journal of Psychotherapy, 38, 190-202.
Westen, D. (1985). Self and socie5: Narczssism, collectivtsm, and the development of moraLs. Cambridge:
Cambridge University Press.
Westen, D. (1988). Transference and information processing. Clinzcal Psychology Reuzeut, 8, 161-179.
Westen, D. (1989). Are “primitive” object relations really preoedipal? Amerzcan Journal of Orthopsy-
chiatry, 59, 331-345.
Westen, D. (in press). Social cognition and object relations. Psycholog& Bulletin.
Westen, D. (1990). Toward a revised theory of borderline object relations: Implications of empirical
research. International Journal of Psycho-Analysis, 71, 661-693.
Westen, D., Lohr, N., Silk, K., Gold, L., & Kerber, K. (1990). Object relations and social cognition in
borderlines, major depressives, and normals: A TAT analysis. Psychological Assessment: A Journal of
Consulting and Clinical Psychology, 2, 355-364.
Westen, D., Ludolph, P., Block, J., Wixom, J., & Wiss, F. C. (1990). Developmental history and object
relations in psychiatrically disturbed adolescent girls. Amerxan Journal of Psychiatry, 147, 1061-1068.
Westen, D., Ludolph, P., Lerner, H., Ruffins, S., & Wiss, F. C. (1990). Object relations in borderline
adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 338-348.
Westen, D., Ludolph, P., Misle, B., Ruffins, S., & Block, J. (1990). Physical and sexual abuse in
adolescent girls with borderline personality disorder. American Journal of Orthopsychiatry, 60, 55-66.
Westen, D., Moses, M. J., Silk, K. R., Lohr, N. E., Cohen, R., & Segal, H. (1989). Quality of depressive
experience in borderline personality disorder and major depression: When depression is not just
depression. Department of Psychology, University of Michigan (unpublished paper).
Winnicott, D. (1971). Playing and reality. New York: Basic Books.
Wixom, J. M. (1988). The depressive experiences of adolescents with borderline personality disorder. Unpub-
lished doctoral dissertation, University of Michigan.
Zanarini, M. C., Gunderson, J. G., & Frankenberg, F. R. (1990). Cognitive features of borderline
personality disorder. Amencan Journal of Psychiatry, 147, 57-63.