TECHNICAL :

CONSIDERATIONS :
IN THE TREATMENT :
: 01-1-0
F. KERNBERG,
M.D.
OF BORDERLINE :
PERSONALITY
ORGANIZATION :

General PrinclIples of Treatment

0 PINION REGARDING THE TREATMENT of borderline pathol-

ogy remains divided. At one extreme are those who would
follow Zetzel’s (1971) approach, which implies that borderline
pathology is related, at least in an important part, to defective
development of ego functions and therefore requires a
supportive psychotherapeutic approach. At the other extreme
are those who would consider a nonmodified psychoanalysis
the optimal treatment for all cases (Segal, 1967). A majority of
clinicians who have worked intensively with borderline
patients have been shifting in recent years, it seems to me,
from a supportive approach inspired by Knight’s earlier work
(1953a, 1953b) to modified psychoanalytic techniques or
psychoanalytic psychotherapy for most patients, while still

Professor of Psychiatry, Cornell hfedical College, and Medical Director, New

York Hospital-Cornell Medical Center, Westchestcr Division, White Plains, N.Y.
Presented at the New York Psychoanalytic Institute and Society, November,
1975, and at the Fall Meeting of the American Psychoanalytic Association,
December, 1975. The author thanks Drs. Jacob Arlow, Leon Shapiro, Martin
Stein, and Arthur Valenstein for their criticisms and suggestions.

795
796 OTTO F. KERNBERG

considering the possibility that some patients may be treated

by nonmodified psychoanalysis from the beginning of
treatment, and others with a modified psychoanalytic
procedure which might gradually evolve into a standard
psychoanalytic situation at advanced stages of the treatment
(Stone, 1954; Greenson, 1970; Jacobson, 1971; Frosch, 1971;
Giovacchini, 1975).
One problem complicating a review of where various
writers stand regarding the issues mentioned is that some of
them, in my opinion, do not sufficiently differentiate
psychoanalysis proper from psychoanalytic psychotherapy;
and many of them, when discussing psychotherapy, imply
such a broad spectrum of psychotherapeutic tools and
techniques that it is hard to know whether they are employing
an expressive or supportive strategy. In attempting to clarify
some of these issues, I shall utilize the following over-all frame
of reference (which should permit the reader to place my
approach within the various alternatives just summarized).
First, I think that, while some borderline patients may
respond to a nonmodified psychoanalytic approach, the vast
majority respond best to a modified psychoanalytic procedure
or psychoanalytic psychotherapy, which I have described in
detail elsewhere (Kernberg, 1968, 1975a). I believe that for
some borderline patients a psychoanalytic approach-stand-
ard or modified-is contraindicated, and these patients do
require a supportive psychotherapy (that is, an approach
based upon a psychoanalytic model of psychotherapy relying
mostly on the supportive techniques outlined by Bibring,
1954; Gill, 1954; and Zetzel, 1971).
Second, I think that psychoanalysis and psychotherapy
should be most carefully differentiated, and I follow Gill
(1954) in this regard; I also believe that the psychoanalytic
psychotherapy I have proposed for borderline patients can
indeed be differentiated from broader psychotherapeutic
approaches which include both expressive or interpretive and
supportive measures, and I will dedicate a good part of this
 TREATMENT OF BORDERLINES 797

paper to spelling out the specific characteristics of my

approach.
Third, I think that much of what appears as “ego
weakness,” in the sense of a defect of these patients, turns out,
under a psychoanalytically based exploration, to reflect con-
flictually determined issues; for example, what first appears as
inability to establish object relations or unavailability of drive
derivatives or lack of affective response or simply lack of
impulse control eventually reflects active defenses against very
intense and primitive object relations in the transference.
Obviously, this conviction underlies my stress on the value of
an interpretive, in contrast to a supportive, approach with
borderline patients. A major additional source for this con-
viction stems from the Psychotherapy Research Project of the
Menninger Foundation (Kernberg et al., 1972), which
revealed, contrary to our initial expectations, that borderline
patients did much better with an interpretive or expressive
approach, and much more poorly with a purely supportive
one.
Fourth, while I believe that, in addition to the effects of
interpretation, there are also therapeutic effects of the
patient-therapist relationship per se which are crucial in the
treatment of borderline conditions -and which include
ego-supportive effects if not techniques- these nonspecific
effects can be best activated in a psychoanalytic atmosphere
that combines the therapist’s technical neutrality with an
interpretive approach. Again, this point will be elaborated
upon in what follows.
The focus of this paper is mostly on patients who would
not be able to undergo and/or benefit from a standard
psychoanalysis, at least not initially or for extended periods of
their treatment. I am strongly convinced, however, that only
psychoanalytic theory and technique and a solid grasp of
normal and pathological psychic development permit the
carrying out of such modified treatment procedures as are
indicated for these patients. Psychoanalytic psychotherapy,
798 O m 0 F. KERNBERC

therefore, is a legitimate and even essential technique for the

psychoanalyst; it broadens the spectrum of psychoanalytic
work and should not be regarded as a diluted or even distorted
analytic approach. I think we need a strict delimitation
between psychoanalysis as such and other modalities, a precise
definition of psychotherapeutic procedures, and an integrative
theoretical and clinical frame for all psychoanalytically
derived psychotherapeutic approaches. Wallerstein (1969) has
raised some fundamental questions regarding these issues and
comprehensively summarized’the literature.
As one effort in this direction I will now attempt to define
“psychoanalytic psychotherapy,” referring here to that limited
approach which is based mostly on clarification and
interpretation and which corresponds to what Bibring (1954)
and Gill (1954)referred to as exploratory psychotherapy. The
implication is that this interpretive approach should be
differentiated sharply from other psychoanalytically derived
procedures. I would tentatively define psychoanalytic
psychotherapy (thus restricted) as the carrying out of
interpretive work, including transference interpretation, in
the context of a treatment setting of technical neutrality
limited only by the need (stemming from the patient’s reality)
to safeguard the patient’s immediate life situation and to
block damaging effects of transference acting out: the setting
up of parameters of techniques (Eider, 1953) is determined
by these requirements. I agree with Brenner (1969) in the
restricted definition of acting out as transference acting out,
and I use the term “transference acting out’’ to lay emphasis
on the restrictive meaning of the term. “Blocking” of acting
out does not refer to moralistic or omnipotent efforts to
control the patient’s life, but to technically required
disruption of self-perpetuating behavior patterns that would
otherwise threaten the patient’s life, the treatment itself, or
produce chronic therapeutic stalemates. Without such
“blocking,” treatment would not be possible. (The dynamic
tension between such measures and the still crucial
 TREATMENT OF BORDERLINES 799

requirements for moving into a direction of technical

neutrality will be dealt with later on.)
Interpretation in psychoanalytic psychotherapy, rather
than being determined by the natural sequence of evolving
transference paradigms (so that a full-fledged transference
neurosis might be systematically worked through), is deter-
mined by the requirements of the dynamic combination of the
predominant transference paradigm, the urgency of immedi-
ate life problems, and the specific over-all goals of the treat-
ment.
Considerations of time and space do not permit detailing
the reasons that underlie my stress on the importance of these
three elements. In brief, I consider the severity of transference
acting out of these patients to imply an additional dimension
of danger to the continuity of the treatment situation, a
danger either nonexistent or much less acute in the ordinary
analytic situation. Acting out in borderline patients often
includes a dimension of magical playing out of the transfer-
ence situation in external reality, as part of an effort to deny
or destroy all reality that goes counter to the transference
wishes; and this needs to be interpreted. With regard to treat-
ment goals, stubborn, long-term treatment stalemates can
often be uncovered only in the context of the patient’s denial
of the passage of time, his treatment of the psychotherapeutic
relationship as an “eternal” one, so that a long-range
transference acting out may become diagnosable only in terms
of comparison of the treatment goals with where the treatment
situation has become immobilized over a long period of time.
Insofar as the psychotherapeutic technique proposed still
requires a technical position of (or persistent movement
toward) neutrality of the analyst and a consistent interpreta-
tion of the transference and other resistances, the tactics of
therapeutic interventions are quite similar to psychoanalysis
proper, although the treatment goals and strategy are
different.
Because this paper focuses on tactical rather than
800 OTTO F. KERNBERG

strategical issues, particularly on the interpretation of the

transference, the technical considerations included may
apply, at least in part, to both a standard psychoanalysis and
psychoanalytic psychotherapy of borderline patients. I am
emphasizing that some borderline patients may indeed be
treated by a nonmodified psychoanalytic procedure. I have
outlined my proposals for the treatment of borderline
personality organization elsewhere (1968, 1975a, 1975c), and
shall only stress here that the essential task is the diagnosis and
resolution of fi rimit ive transference paradigms.
Particularly in the early stages of treatment, the trans-
ference is either characterized predominantly by overwhelm-
ing chaos, meaninglessness, or emptiness, or it is consciously
suppressed or distorted. This usually results from the pre-
dominance of “primitive transferences,” that is, the activation
in the transference of “part object” relations-or units of early
self and object images and the primitive affects linking
them-not characteristic of internal object relations of
neurotic patients and normal people. The transference
reflects a multitude of internal object relations of dissociated
(or split-off) aspects of the self and highly distorted and
fantastic dissociated (or split-off) object representations.
The strategical aim in working through the transference
is to resolve these primitive dissociations of self and internal-
ized objects and thus to transform primitive transferences into
higher level or integrated transference reactions, more
realistic and more related to real childhood experiences
(1975~).Obviously, this requires intensive, long-term treat-
ment along the lines I have suggested, usually not less than
three sessions a week over years of treatment. First, the
dissociated or generally fragmented aspects of the patient’s
intrapsychic conflicts are gradually integrated into significant
units of primitive internalized object relations. Second, each
unit (constituted of a certain self image, a certain object
image, and a major affect disposition linking these) then needs
to be clarified as it becomes activated in the transference,
 TREATMENT OF BORDERLINES 801

including the alternation of reciprocal self and object

re-enactments. Third, when these units can be interpreted
and integrated with other related or contradictory units-
particularly when libidinally invested and aggressively
invested units can be integrated-the process of working
through of the transference and of the resolution of primitive
constellations of defensive operations characteristic of border-
line conditions has begun.
When such a resolution of primitive transferences has
occurred, the integrative affect dispositions that now emerge
reflect more coherent and differentiated drive derivatives.
The integrative object images now reflect more realistic
parental images as perceived in early childhood.

Two Clinical Illustrations

Case 1
A 30-year-old architect had begun psychoanalysis four months
earlier with the following major symptoms: chronic feelings of
depression, emotional inhibition in group situations, fear-
fulness of being criticized, and a general feeling of lack of
authenticity. The diagnosis was depressive personality with
strong paranoid features, and possibly borderline personality
organization. In spite of severe pathology of object relations
and a strong predominance of primitive-particularly
paranoid- defenses, there was sufficient evidence of non-
specific ego strength to warrant a standard psychoanalysis.
From the beginning of his analysis, the patient rapidly
shifted from one subject to another during the sessions. What
he talked about seemed important, but the manner of
presentation seemed to me inauthentic. Either the patient
stayed with one subject but talked in an intellectualized way
about it, or the subject matter sounded authentic enough, but
would only appear for fleeting moments in the middle of other
material. The patient seemed distant, almost strangely aloof,
802 OTTO F. KERNBERC

and yet had moments of great emotional intensity-gone too

fast to be fully understood. He was constantly preoccupied
with any possible criticism from me and with the possibility
that I might not understand him or might show him a “phony”
friendliness.
In one session, the patient began by remembering the
impossible situation he felt himself in when he was about ten
years old. At this time his mother, a chronically depressed
alcoholic, failed to display any warmth or even the merest
interest in him. Her interest occurred only when she was under
the effect of alcohol, and he experienced it as false and
embarrassing; when she was sober, she was cold and distant.
Another subject then appeared in his associations, namely, his
feelings of dread and guilt over the incapacitating illness of his
cousin, who was reduced to almost complete immobility
without any hope of improvement. This cousin had been
competing with the patient throughout their adolescence, and
the patient was quite conscious of his sense of triumph-and
yet deep guilt-about this illness. In the past, the patient had
experienced fear of my criticizing him because of his sense of
triumph over his cousin, and he had once felt irritated at my
acknowledgment of his feelings of guilt toward his cousin: the
patient felt that my understanding was an expression of phony
concern on my part.
In the hour under discussion, the following developments
took place in rapid succession: as I was thinking that the
patient was irritated at what he saw as my concern over his
conflicts with his cousin because he experienced me as being as
phony as his mother was when she was seemingly interested in
him under the influence of alcohol, the patient said that he
was sure I was speculating about this (indeed, guessing my
thoughts) and that it reflected my complete lack of
understanding of what was going on in the analytic situation.
He reminded me that the day before he had accused me of
being very critical although attempting to hide that criticism.
(I had said something not critical to him which he construed
 TREATMENT OF BORDERLINES 803

as criticism in response to his telling me that he had been

smoking pot.) He then said that he really wouldn’t mind if I
did not understand or were indifferent to him, but he resented
what he saw as the rigid, petty, provincial quality of my
criticism, a reflection of my rigidity and puritanism related to
being part of the psychoanalytic establishment.
He then thought of his wife, a business woman whose
functions took her out of town for several days at a time, and
speculated if she would ever have a n affair while being away.
In contrast to the earlier associations about his mother and his
cousin, these latter ones seemed to contain more direct,
strong, although fleeting, emotions. A silence ensued, and the
patient said that he could imagine my clicking away various
interpretations and my trying to hide from him my insecurity.
My conviction gradually strengthened that the most
important thing going on was the patient’s almost desperate
need to fill the hour with material that would seem important
to me-while he would remain “one step ahead” and attempt
to keep the control over my thinking and feeling. Something
very actively going on in the transference was being
submerged by the patient’s attempt at “thought control.” I
finally told him I thought he was trying to bring up a number
of matters that were meaningful, indeed, but that now served
the purpose of preventing the emergence of a more dreaded
emotional experience regarding me.
He then became anxious, started to breathe rapidly, and
said that something was terrorizing him but he couldn’t
explain it. He had had moments when he experienced terror
when his mother-who had died during the patient’s
preadolescent years-looked at him shortly before the time of
her death. He then remembered the administrator of the
building he lived in, whom he thought dishonest and trying to
control people in surreptitious ways (and, if that were not
possible, by force); the patient said that he felt like
terminating his analysis right then and there. It turned out
that, at one level, the patient was identifying me with the sick
804 OTTO F. KERNBERG

cousin and his dying mother, while he identified himself with

the controlling, sadistic, dishonest administrator. This
relationship, in turn, was a defense against the opposite one:
on a deeper level, he was the potential victim and I the
persecutor.
It eventually became clearer to him that he was
attempting to escape from an experience of undefinable dread
in the hour related to the image of me as somebody who would
hatefully attack him after pretending either indifference or
phony interest. What I wish to stress is the uncanny nature of
this emotional experience, the overwhelming fear and dread
of my aggression, the fantastic nature of the image of me, and
the difference between this essential primitive aspect of the
transference, on the one hand, and the various more
sophisticated transference dispositions related to his guilt and
aggression toward his mother, his competition with the cousin,
and the jealousy of his wife, on the other.
The rapid shift of his associations, the “invasive” quality
of his focus on my thinking, his disqualification of my
comments, and the rapid shift of his feelings and attitudes, all
brought about a chaotic combination of material in the
middle of which it was hard to say what object relation
activated in the transference was predominant. The
reconstruction of that dominant object relation on the basis of
the examination of the total situation in the hour-including
my emotional reactions to the material- permitted clarifica-
tion of the nature of the primitive transference activated in the
psychoanalytic situation. This case, it needs to be stressed, was
still a rather “standard” analytic case; the next one is a more
typically regressed borderline patient.

Case 2

A graduate student in her late twenties had begun

psychoanalytic psychotherapy (three sessions a week) a few
weeks earlier because of a severe depressive reaction-with
 TREATMENT OF BORDERLINES 805

weight loss and suicidal ideation, alcoholism, and a general

breakdown in her functioning at school, in her social life, and
in her love relation with her boy friend. The diagnosis was
infantile personality with borderline features, severe depres-
sive reaction, and symptomatic alcoholism. This treatment
was carried out in “face-to-face”sessions (for reasons I have
mentioned in earlier work [Kernberg, 1975a], I only utilize the
couch in cases of standard psychoanalysis, and never in cases
of psychoanalytic psychotherapy as defined, or with any
modality of psychotherapeutic treatment). In the beginning of
the treatment, I had established a number of conditions under
which I would be willing to see her on an out-patient basis. If
she could not fulfill these conditions, I would see her in
psychotherapy and she would remain in the hospital until she
became ready to fulfill these conditions. Hospitalization had
been suggested by other psychiatrists who had seen the
patient, and I, too, had contemplated it as an alternative, in
case she proved unable to take responsibility for her
immediate functioning in reality.
Insofar as I made these preconditions, one might say that
a selective process took place and that only a limited range of
borderline patients would accept and be able to undergo the
kind of treatment approach proposed. However, it must be
pointed out that this patient had available as an alternative
the possibility of short-term (or long-term) hospitalization,
and that, if needed, I would expect hospitalization at a
minimum to develop the patient’s capacity to take
responsibility for such functions as are indispensable in any
case for out-patient treatment. The implication is that the
opposite approach, namely, to accept out-patient treatment
on a less than realistic basis, would bring about a more
complicated and potentially disastrous course of treatment. In
other words, a therapist cannot do justice to his patient’s needs
if the therapist does not have the minimum requirements and
freedom for full deployment of his special technical knowledge
and capabilities.
806 OTTO F. KERNBERG

This patient had committed herself to stop drinking, not

to act on suicidal impulses while discussing them openly with
me (if and when they occurred), and to maintain a minimum
weight by eating sufficiently, regardless of her mood and
appetite. A psychiatric social worker was beginning an
evaluation of the patient’s total social situation, including her
relation with her parents, who lived in a different town, and
was available to the patient for any suggestion and advice, if
needed, regarding any problem in her daily life. Our
understanding was that the social worker would convey full
information about the patient to me, and that I would
communicate to the social worker only that information I
considered crucial and that the patient had explicitly
authorized me to transmit. My psychotherapeutic approach
was essentially psychoanalytic, and I attempted to maintain a
consistent position of technical neutrality (further discussion
of which will follow).
In the session to be described, the patient looked haggard
and distraught; it had been raining, and she had not taken
precautions to dress correspondingly, so that she came without
raincoat or umbrella, drenching wet, the thinness of her body
showing under the wet clothes. Her blouse and pants were
dirty, and she looked somewhat disheveled. She began talking
immediately about a difficult test she would have to take at
school that she was afraid she would not pass. She then talked
about a serious fight with her boy friend; she felt jealous
because of his interest in another woman, a former girl friend
of his whom, the patient had discovered, he had met secretly.
She also expressed concern over her parents’ sending her the
monthly check (which produced in me the fantasy that she
wanted to reassure me: she would be paying for her treatment,
and therefore, even if I were fed up with her otherwise, it
would still be in my interest to continue seeing her in spite of
her “unlikable” nature).
The patient’s ffow of verbal communication seemed to be
disrupted by a variety of nonverbal behaviors. Sporadically,
 TREATMENT OF BORDERLINES 807

she fell silent and looked at me with a searching, distrustful,

and somewhat withdrawn expression; there were moments of
inappropriate, artificial gaiety, and a forced laughter, which
conveyed to me her conscious efforts to control our
interaction. She became “confused” in telling me where she
had met her boy friend and the other woman-which made
me wonder whether she had been drinking without daring to
let me know about it. Information from the psychiatric social
worker came to my mind: the patient had casually observed to
her that she had had stomach aches and had been vomiting in
recent days, and the social worker wondered whether the
patient would not need another medical check-up.
I now oscillated between moments of concern and strong
urges to express this concern to the patient in terms of t3e
deterioration of her physical appearance, her health, and the
question whether she would really be able to maintain the
out-patient treatment setting we had agreed upon. I was also
tempted to confront her with those aspects in her behavior
which made me wonder whether she was telling me the truth.
I felt an underlying fearfulness in her, and an experience of
me as a potentially critical and inflexible parental figure who
would scold her for not being truthful or for behaving poorly.
(All this was superimposed on my feeling that she desperately
wanted me to take over and run her life.) I now became aware
of a growing sense of impatience in myself, a combination of
worry for the patient and yet an irritation that the treatment
program as set up was falling apart, and that inordinate
demands were being made on the psychiatric social worker
and on me to change the treatment arrangements, mobilize
her parents, and protect her against the impending threat of
dismissal from school (another reality aspect which had
brought her into treatment).
I finally felt that the predominant human relationship
enacted at that moment was of a frightened little girl who
wanted a powerful parental figure (the particular sexual
identity of whom was irrelevant) to take over and protect her
808 OTTO F. KERNBERG

from pain and fear, from suffering in general. At the same

time, I thought she hated that parental figure because such a
taking over could only be forced by extreme circumstances of
the patient’s suffering and not by natural concern, love, and
dedication to her. And she was afraid of a retaliatory attack
from that needed and yet resented parental figure because she
projected her own angry demands onto it. Therefore, I felt,
she had to escape from that dreaded relationship, perhaps
drink herself into oblivion, and create a situation of chaos in
which she would be rescued without having to acknowledge or
emotionally relate to the rescuer as an enemy to fight off.
I now said to the patient that, on the basis of what she
seemed to be communicating, I had the feeling that she was
expressing contradictory wishes: she wanted to reassure me
that she was still in control over her life, while at the same time
she was conveying almost dramatically that things were falling
apart, that she was unable to handle her life, and that she was
running the risk of illness, expulsion from school, and loss of
her boy friend. I added that, if I were correct, whatever I
might do under these circumstances would be disastrous to
her: on the one hand, if I explored further whether she was
really able to handle her immediate life situation, she would
experience this as an attack. For example, if I raised the
question whether, under these circumstances; she felt able to
continue not to drink, it would be a “cross examination”
revealing my basic harshness and suspiciousness of her. If, on
the other hand, I sympathetically listened without raising any
of these questions, it would be like an indifferent, callous
expression of a psychiatrist mostly worried about whether he
would be paid by the patient’s parents. In either case, she
could expect only suffering and disappointment from my
potential reactions to her.
The patient replied that she had been worried because
she had been drinking and was afraid I would hospitalize her
if I knew. She had also thought that, because of her parents’
lateness in sending her the money, I might be worried about
 TREATMENT OF BORDERLINES 809

receiving my fee, and she had been angry at her parents and
embarrassed about this lateness. She said she felt completely
hopeless about herself, and didn’t know how this psycho-
therapy could help her. I said I thought she was not so much
worried about whether the psychotherapy would help her, but
whether I was genuinely interested in her, or just concerned
about getting paid for the sessions. I also said that I wondered
whether she felt that the only way of obtaining anything from
me was by forcing it from me, extracting it, so to speak, by
presenting herself as if she were a completely helpless human
wreck. I added that, under the circumstances, any help from
me would be like the irritated, angry reaction of a parent who
would prefer not to be bothered, but had no alternative but to
take care of an unwanted child.
The patient burst into tears, said how desolate she felt
because her boy friend was leaving her, adding with real
feeling that she felt she had no right to be helped by me, there
were many people like her, and only because of the fortunate
circumstance that her parents had money was she able to
afford treatment that would otherwise be unavailable. Why
should she be treated, when there was so much suffering in the
world? It was better to give up . ..
At this moment, I felt that a change had occurred in the
hour: now her verbal communications and her behavior
coincided, she had become open in telling me about what had
actually happened during the last two days, there was an
awareness of and worry over her sense of failure, mixed with a
strange sense of relief for failing because she felt she did not
deserve better. I felt that the patient was becoming concerned
for herself in the process of this interchange, that the pressure
on me for taking over was decreasing, and that she was
beginning to feel guilty for having failed to keep her part of
our agreement, while beginning to understand that this
failure was an expression of her sense that she did not deserve
to be helped.
What I wish to stress is that the situation now became a
810 OTTO F. KERNBERG

more coherent transference situation in which a masochistic

character pattern was expressed in a mostly pregenital,
conflictually dependent relation to a frustrating parental
image. The intersystemic conflict between superego features
(unconscious guilt) and the dependency conflictually
expressed by her ego now reflected a rather typical neurotic
object relation in the transference. This development, then,
needs to be contrasted with the earlier, chaotic, contradictory
manifestations of implicit suspicion, projected anger,
aggressive demandingness, concealment, and withdrawal. I
would also stress that the reduction of the chaotic transference
manifestations into the predominant object relation expressed
in the transference permitted the full exploration of the
transference and life situations and the activation of ego
resources that made it unnecessary for the analyst to intervene
on the patient’s behalf. In other words, the approach from a
position of technical neutrality permitted a strengthening, if
only temporarily, of the patient’s ego and a fostering of her
capacity to combine understanding in the hour with an
increasing sense of responsibility for her life outside the
treat men t hours.

Further Considerations Regarding the Charactertitics of

Primitive Transference and Their Interpretation
What follows are some technical considerations that seem to
me of particular relevance for the psychoanalyst treating
borderline patients with psychoanalysis or psychoanalytic
psychotherapy, and an attempt to elucidate further some of
the implications of the general principles of treatment
mentioned earlier in this paper.
Borderline patients charrcteristically present primitive
transferences. In practice, this means that the activation of an
ordinary transference is replaced by an impulse derivative of a
more primitive nature, often by peculiar condensations of
aggressive and libidinal drive derivatives reflected in diffuse,
 TREATMENT OF BORDERLINES 81 1

overwhelming affect states. Instead of a definite projection of

a certain infantile object image onto the analyst while the
patient reactivates aspects of the infantile self in that relation,
borderline patients reactivate dissociated self and object
images, with rapid oscillation of the projection of either self or
object component of that relation onto the analyst. The
projected object image is often highly unrealistic or strange,
and does not reflect the repetition of a real infantile or child-
hood experience; the patient’s self-experience in regard to this
object image projected onto the analyst also reflects a bizarre,
strange experience of the self that is not linked to or integrated
with a more global infantile self activated in the transference.
In other words, the ordinary object relation in the
neurotic transference is replaced by a fantastic relationship in
primitive transferences, within which subject and object are
easily interchangeable and remarkably unrelated to the usual
or predominant characteristics of the particular infantile self
or parental figure involved. The diffuse, overwhelming,
nonmodulated characteristics of the affects involved are
particularly striking in regard to the intensity of anxiety,
which frequently borders on panic; the affects reflecting
dependency, love, and aggression are dissociated from other,
contradictory feelings toward the same object that may be
present in the transference only minutes or hours before or
after that particular affect is present. In short, affects, self
images, and object images shift rapidly and chaotically, and
the reconstruction of more complex and integrated past object
relations requires that the analyst try to integrate the mutually
dissociated transference aspects into what eventually reflects
more realistic childhood experiences.
What complicates the picture further is that, not only is
there a rapid escalation or change of transference dispositions
(which gives an over-all chaotic picture to the transference
developments), but the very nature of primitive defensive
operations- particularly severe forms of splitting (leading to
fragmentation of all emotional experience), omnipotent
812 OTTO F. KERNBERG

control, and devaluation, etc. -all bring about a general

deterioration or destruction of human experience in the
hours, so that even the dissociated primitive internalized
object relations are at first unavailable and may need to be
reactivated or reconstructed by means of the interpretation of
these primitive defenses. In other words, often the analyst’s
task is first to work through meaninglessness, generalized
dispersal of emotions, and paralyzing emptiness or distortions
in the hours, in order to permit the full development, later, of
the primitive transference manifestations mentioned. Only
later still, and very gradually, may these chaotic transference
dispositions be integrated into broader transference paradigms
of the usual neurotic kind.
The implications for the technique of diagnosing and
interpreting the transference are that the analyst has to focus
sharply on manifestations of emptiness, meaninglessness,
distortion, and control, which stand in the way of the full
deployment of primitive transference dispositions. Once a
human relationship has been activated in the transference so
that the experience of the analyst is of being part of a
primitive, overwhelming, emotionally charged relationship,
the analyst has to diagnose and verbalize the nature of this
relation and to define the self and object component in it. The
questions one may ask regarding any transference mani-
festation, namely, “Who is saying or doing what to whom now
and why?” (Heimann, 1956) will have to be answered, under
typical circumstances with borderline patients, in terms of
a highly unrealistic, primitive interaction between an aspect
of the self and fantastically distorted object representations. It
is only after such primitive relations have been diagnosed,
clarified, and spelled out repeatedly-while their self and
object aspects are exchanged, as it were, between patient and
analyst- that the defensive quality of the mutual dissociation
of these primitive transferences can be diagnosed and
interpreted and a n integration of self and object components
can occur.
 TREATMENT OF BORDERLINES 813

The characteristics of the transferences of borderline

patients make genetic reconstructions very difficult or
impossible. Insofar as the primitive transference dispositions
reflect fantastic, unreal, internal relations of dissociated
aspects of the self with “part objects,” what is reconstructed is
really an internal world of object relations expressed largely in
primary-process thinking, with qualities of timelessness and
condensation of various stages of psychosexual development,
all of which make genetic reconstructions highly speculative,
to put it mildly. It is only at later stages of treatment, when
more realistic childhood experiences are reactivated in more
advanced levels of transference manifestations, that efforts to
reconstruct genetically significant childhood conflicts become
possible and effective. It is a typical characteristic of the
clinical study of borderline patients that the initial history
reveals little of what later on turn out to be the main
transference paradigms in the treatment, and the differences
between initial information and genetic reconstructions are
much sharper than is the case with the average neurotic
patient.
Therefore, the analyst has to tolerate a state of ignorance
and uncertainty about the genetic continuity of the material
activated in the transference that exceeds by far those of the
ordinary psychoanalytic treatment. What makes the situation
even more difficult is the very intensity of primitive trans-
ferences and the weakening of reality testing in the area of
projection onto the therapist of primitive impulses and self
and/or object images. Primitive types of projection, particu-
larly projective identification, bring about a confusion, on the
part of the patient, of present interaction with past internal
object relations. The reason for this is the patient’s confusion
of what is “inside” and what is “outside” under these
circumstances, so that ego boundaries, reality testing, and
secondary-process thinking are all weakened. Any premature
efforts to achieve immediate genetic reconstructions of such
fantastic transferences may induce in the patient the
814 OTTO F. KERNBERG

conviction that the fantastic internal object relation was at one

time-and is again now-a real one, and past and present are
just the same. In other words, premature genetic reconstruc-
tion may foster transference psychosis.
The nature of primitive transferences in borderline
patients presents certain technical problems and dangers. The
analyst may be tempted to interpret these transferences
directly, as if they reflected the actual, earliest, or most
primitive human experiences: he might even go so far as to
interpret them as a genetic reconstruction of the first few years
or even the first few months of life, thus confusing or
condensing primitive fantasy and actual earliest development,
characteristic of some Kleinian work. The Kleinian approach
combines, in my opinion, two errors: first, the mistaking of
the primitive, bizarre intrapsychic elaboration of psychic
experience with actual developmental features, and second,
the telescoping of complex, slowly developing structural
organization of internalized object relations with assumed
(and highly questionable) developments in the first few
months of life.
A second danger can result from assuming a simplistic
ego-psychological approach. The intense activation of affects
in a patient with little capacity for observing what he is
experiencing may lead the analyst to focus on ego functioning,
to the neglect of the object-relations implications of what is
activated in the transference. The analyst may focus, for
example, on the patient’s difficulty in experiencing or
expressing his feelings or overcoming silence, his tendency to
impulsive actions, or his temporary loss of logical clarity,
instead of on the total primitive human interaction (or the
defenses against it) activated in the transference. A mistake in
the opposite direction would be to interpret the object relation
“in depth” without paying sufficient heed to the patient’s ego
functions. A further danger is of focusing exclusively on the
“here and now,” in the context of conceptualizing the
transference as a corrective emotional “encounter” and
 TREATMENT OF BORDERLINES 815

neglecting the task of gradually integrating self and object

images into more realistic internalized object relations and
advanced types of transference that will permit more realistic
genetic reconstructions. Here, the analyst unwillingly or
unwittingly may contribute to the stability of primitive
transferences (as the treatment replaces life), thus interfering
with the patient’s ego growth.
To give simultaneous attention to the “here and now”
and to the underlying primitive internalized object relations
activated in the transference so that what is on the surface and
what is deepest are integrated into human experiences of
ever-growing complexity is implicit in the technical approach
I have proposed. In this process, whatever remnant the patient
has of a capacity for self-observation and autonomous work on
his problems must be explored, highlighted, and reinforced,
so that attention is given to the patient’s ego functioning,
particularly to his self-observing ego function, hand in hand
with the clarification and verbalization of primitive object
relations reflected in his conscious and unconscious fantasies.
Fenichel’s (1941) general rule of interpretation- to
proceed from surface to depth-certainly applies to
borderline patients. It is helpful if we first share our
observations with the patient, stimulate him to integrate them
a step beyond what is immediately observable, and only
interpret beyond his own awareness when it is clear that he
cannot do so himself. Further, whenever we interpret beyond
the patient’s awareness of the transference situation, we
should include in our interpretation the reasons for his
unawareness. Inasmuch as primitive transference dispositions
imply a rapid shift to a deep level of experience, the analyst
working with borderline patients must be prepared to shift his
focus from the “here and now” to the fantasied object relation
activated in the transference- one that often includes
bizarre and primitive characteristics which the analyst has to
dare to make verbally explicit as far as his understanding
permits. Moreover, the analyst must also be alert to the
816 OTTO F. KERNBERG

danger that the patient interpret what he has said as a magical

statement, derived from a magical understanding, rather than
a realistic putting-together of what the patient has communi-
cated to him.
Integrative aspects of the interpretations, therefore, in-
clude consistent interpretation of surface and depth. The same
procedure applies to the patient’s communications, so that
when apparently “deep” material comes up in the patient’s
communications it is important to clarify, first, to what ex-
tent the patient is expressing an emotional experience, an
intellectual speculation, a fantasy, or a delusional conviction.
One question the analyst often has to ask himself in the
treatment of borderline patients, “Should I now clarify reality,
or should I now interpret in depth?”, can usually be answered
by evaluating the patient’s reality testing at the moment, his
capacity for self-observation, and the disorganizing effects of
primitive defensive operations in the transference. Ideally,
clarification of reality and interpretation in depth should be
integrated, but that is often not immediately possible.
Surprisingly enough, as one’s experience with this kind of
treatment of borderline patients grows, one finds more and
more that what seemed at first a simple manifestation of ego
weakness or ego defect turns out to be the effect of very
specific, active, primitive defensive operations directed
against full awareness (on the part of the patient) of a
dissociated transference relation reflecting intrapsychic
conflicts. In other words, this treatment approach permits us
to diagnose areas of ego weakness, to evaluate the
ego-weakening effects of primitive defensive operations and of
dissociated internalized object relations, and to foster ego
growth by essentially interpretive means. Clarifications of
reality made by the analyst often subsequently turn out to
have been an unnecessary support feeding into a certain
transference situation. Whenever this is the case, it is very
helpful to interpret to the patient how an apparently necessary
clarification by the analyst was actually not necessary at all.
 TREATMENT OF BORDERLINES 817

Often the question arises of whether to further clarify the

reality of the patient, or whether to interpret the meaning of
the patient’s distortion of reality. If one proceeds from the
surface to depth, first testing the limits of the patient’s
understanding, and then interpreting the defensive aspects of
the patient’s lack of awareness of an appropriate perception of
(or reaction to) reality, one can usually resolve this apparent
alternative in a basically analytic fashion. The danger always
exists that the patient will interpret our interpretation of his
defensive denial of reality as a subtle attempt to influence
him. Therefore, in addition to our introspective evaluation of
whether the patient may be right in this regard, this distortion
of our interpretation has to be interpreted as well. In essence,
technical neutrality, interpretation of the transference, the
analyst’s introspective exploration of his countertransference,
and focus on the patient’s perceptions in the hour are all
intimately linked technical tools.
A related problem of the analyst working with borderline
conditions is, to what extent he should intervene quickly in an
interpretive fashion, or to what extent he should wait until the
patient is ready to do further work on his own. In general,
once operational understanding has been achieved, I think
few advantages are gained from simply waiting. Insofar as
primitive transferences are activated rapidly in the hours and
tend to perpetuate themselves in a repetition compulsion
which often defies long-term interpretative work, there is an
advantage to interpreting the material fully as soon as it is
clear enough and whenever a certain transference disposition
becomes a predominant transference resistance. There are
borderline patients who activate one kind of transference
pattern in an endless repetition over many months and years,
and an early interpretive stance may not only save much time,
but protect the patient from destructive acting out. At the
same time, an interpretive approach that deals rapidly with
the developing transference resistances does not imply a
bypassing of the patient’s own capacity for self-observation: I
818 OTTO F. KERNBERG

wish to stress again that interpretation in depth should include

an ongoing evaluation of the patient’s capacity for self-obser-
vation and never justify the patient’s hopes (or fears) of a
magical relation with an omnipotent therapist. Very often,
simply waiting for the patient to improve in his self-observing
capacities is of little usefulness and creates the danger of
bringing about chronic countertransference distortions which
gradually undermine the analyst’s position of technical
neutrality.
I have mentioned elsewhere (1975~)the need to rapidly
deepen the level of interpretation when the patient begins to
act out. Although there is a risk of interpreting beyond the
level of emotional understanding reached by the patient at
that point-and a risk, therefore, that the interpretation be
either rejected or incorporated in intellectualized or magical
ways-the focus on the patient’s relation to the interpretation
will make it possible to correct such potential misfirings of
quick interpretations of the transference at such times.
Another important aspect of the analysis of the
transference of borderline conditions is the tenuous nature of
the therapeutic alliance as compared with that of ordinary
neurotic cases in a standard psychoanalysis. The therapeutic
alliance (Greenson, 1965; Zetzel, 1956) links the analyst as
such-a professional in a special work relation with the
patient-and the observing part of the patient’s ego, however
small or limited, in the treatment situation. Insofar as there is
a sufficient observing part of the patient’s ego in the ordinary
neurotic case, the therapeutic alliance or, one might say, the
task-oriented alliance between the patient’s observing ego and
the analyst is a given, ordinarily not requiring too much
attention. However, even under ideal circumstances in the
typical psychoanalytic case, there are times when the intensity
of the transference relation threatens to overwhelm
temporarily the patient’s observing ego, and it may be
necessary to focus the patient’s attention on that complica-
tion. In contrast, the observing ego in the typical borderline
 TREATMENT OF BORDERLINES 819

case is so limited and frail that the question has been raised, to
what extent there is a need to focus attention strongly,
consistently on the therapeutic alliance: Are there times when
the analyst needs to take active measures to strengthen the
therapeutic alliance by reality-oriented, supportive comments
or by providing the patient with information regarding the
analyst’s reality aspects?
In my opinion, to focus on the defensive use or nature of
the distortions of the patient’s perceptions, and, particularly,
on his distortions of the analyst’s interpretations, are the best
means of strengthening the patient’s observing ego without
shifting from an essentially analytic model. The various
aspects of the real relation between analyst and patient are, it
seems to me, a nonspecific, potentially therapeutic aspect of
the treatment in all cases. However, this aspect of the total
treatment relationship is rather limited in the ordinary
psychoanalytic case, and the systematic analysis of all
transference paradigms in the unfolding transference neurosis
should lead to a systematic working through of the patient’s
efforts to use the therapeutic relationship as a parental
function in the transference. In other words, the nonspecific
supportive implications of the real aspects of the relationship
in terms of the patient’s unfolding transference will auto-
matically become part of the analytic work and remain in the
background. In borderline patients, on the other hand, the
nonspecific, “real” human relationship reflected in the thcra-
peutic alliance may constitute an important corrective
emotional experience, not in the sense of the therapist’s
adopting an active, manipulative stance, but in the sense of
the normally gratifying nature of such a positive human and
working relation, which often goes far beyond anything the
patient had previously experienced. Insofar as chronically
traumatizing or frustrating circumstances of early develop-
ment and, therefore, of the patient’s former relations with the
real parents, are an important aspect of the genetic and
historical background of borderline patients, the real relation
820 OTTO F. KERNBERG

with the therapist may carry out parental functions the patient
has never had before.
In addition, the therapist’s being available to absorb,
organize, and transform the patient’s chaotic intrapsychic
experience (which the therapist first attempts to clarify in his
own mind and then reflects back to the patient as part of his
interpretive comments) does provide cognitive functions which
the ordinary neurotic patient undergoing a standard
psychoanalysis is expected to carry out for himself. In other
words, the therapist does provide auxiliary cognitive ego
functions for the borderline patient in addition to the implicit
reassurance given by his ability to withstand and not be
destroyed by the patient’s aggression, to the therapist’s
neither falling apart nor retaliating, to his maintaining a
general attitude of concern and being emotionally available to
the patient. Others have .stressed the importance of the
therapist’s “mothering” functions with borderline patients
(Little, 1958, 1960; Winnicott, 1958). These aspects of the
therapeutic relationship undoubtedly play an important role
in the psychoanalytic psychotherapy of borderline patients,
and, as long as they occur within a setting of technical
neutrality, constitute a legitimate use of the psychotherapeutic
relationship. This use has to be differentiated from the
patient’s intense transference demands that the analyst gratify
those needs that were previously frustrated, and carry out
active parental functions, thus abandoning the position of
technical neutrality and increasing the supportive aspects of
the therapeutic relationship. I cannot emphasize strongly
enough the need for the psychoanalyst working with
borderline patients to carefully analyze all these attempts on
the patient’s part. What really strengthens the patient’s ego is
not the gratification of needs in the “here and now” that were
denied in the “there and then,” but a coming to terms with
past frustrations and limitations in the context of an
understanding of the pathological reactions, impulses, and
defenses that were activated under those past traumatic
 TREATMENT OF BORDERLINES 821

circumstances and contributed importantly to the develop-

ment and fixation of ego weakness.

Technical Neutrality and Interpretation

I have stressed (1968, 1975a) how crucial it is that the psy-

chotherapist of the borderline patient remain in a position
of technical neutrality-equidistant from external reality, the
patient’s superego, his instinctual needs, and his acting (in
contrast to observing) ego (A. Freud, 1936). It is necessary to
set limits for many borderline patients so that they do not act
out in ways that threaten their treatment or safety. Sometimes
the therapist has to spell out certain conditions the patient
must meet in order for out-patient psychoanalytic psycho-
therapy to proceed. The setting up of such conditions for
treatment represents, of course, the setting up of parameters
of technique (Eissler, 1953). By the same token, the setting up
of such parameters implies a reduction of the position of
technical neutrality, and, beyond a certain point, it is
questionable whether a standard psychoanalysis can be
carried out. Even when the treatment is psychoanalytic
psychotherapy rather than analysis proper, if the therapist
remains constantly vigilant to his deviations from technical
neutrality, it will help him to evaluate the extent to which
transference acting out is occurring, and the extent to which
chronic countertransference distortions are complicating the
treatment situation.
It may be helpful to stress that, as Freud pointed out
(1963), technical neutrality does not mean “listless indiffer-
ence,” or lack of spontaneity and natural warmth. Neutrality
implies, rather, a sufficient degree of objectivity combined
with a n authentic concern for the patient. It also implies a
relative degree of freedom from impingement of general
theoretical formulations that would interfere with his
immediate attention to the patient, and from pressures of any
822 OTTO F. KERNBERG

kind to move, push, direct, or coach the patient into any

particular direction. Technical neutrality, thus conceived,
guarantees not only the analyst’s freedom to carry out
psychoanalytic work, but maximal protection of the patient’s
autonomy and independence and his capacity to carry out
work on his own (in contrast to acting out excessive
dependency in the transference).
In the treatment of borderline patients, neutrality
constitutes an “ideal” but ever transitory situation, which is
threatened continuously from various sources. The danger of
acting out and reality conflicts of borderline patients which
threaten their well-being, the treatment itself, or even their
life, may induce in the analyst an urge to act rather than
interpret. The effect of primitive defensive operations,
particularly of projective identification, is not only to attribute
a certain mental disposition to the analyst, but to induce in
him a certain emotional disposition which complements the
patient’s own affective state, an urge to act in a certain
direction which complements the transference needs. Neutral-
ity is thus challenged or threatened, although by the same
token, every momentary threat to or deviation from technical
neutrality imparts important transference information.
The primitive nature of the transference activated in
borderline patients leads the analyst, as he strives to empathize
with the patient, to whatever capacity he may have for
awareness of primitive emotional reactions within himself.
This is reinforced by the patient’s nonverbal behavior,
especially those aspects of it that attempt to control the
analyst, to impose on him, so to speak, the role assigned to the
self or to an object image within the primitive activated
transference. We probably still do not know enough about
how one person’s behavior may induce emotional and
behavioral reactions in another. The analyst’s emotional
empathy, his creative use of evenly hovering attention-a
function akin to daydreaming-and the direct impact of
behavioral perception, all combine to bring about a
 TREATMENT OF BORDERLINES 823

temporary regressive reaction that permits him to identify

with the patient’s primitive levels of functioning.
In order to maintain an optimal degree of inner freedom
for exploring his own emotional reactions and fantasy
formations in connection with the patient’s material, the
analyst who treats borderline patients must be particularly
concerned that he intervene only when he has again reached a
technically neutral position. It is especially important to
maintain a consistent attitude of “abstinence”-in the sense of
not giving in to the patitnt’s demands for transference
gratification, and rather to interpret fully and consistently
these transference demands. The analyst’s humanity, warmth,
and concern will come through naturally in his ongoing
attention to and work with the patient’s difficulties in the
transference, and in his ability to absorb and yet not react to
the onslaught of the demands stemming from primitive
dependent, sexual and aggressive needs.
In short, technical neutrality, attention to abstinence in
the transference, preservation of the internal freedom for
’

analyzing the transference, and, introspectively only, the

countertransference components of the analyst’s emotional
reaction are intimately linked aspects of the over-all technical
approach to borderline patients.

Psychoanalysis or Psychoanalytic Psychotherapy?

The analyst’s position of technical neutrality in the treatment
of borderline patients may be helped by an early, informed,
and well-considered decision of whether the patient will be
treated with standard psychoanalysis or whether modified,
psychoanalytic psychotherapy will be the treatment of choice.
The implication of such an early decision is that a degree of
ego weakness that would contraindicate a psychoanalysis
proper, or a potential for severe acting out that would
interfere with the maintenance of a standard psychoanalytic
situation can be evaluated as part of the diagnostic process,
824 OTTO F. KERNBERG

and one can thus prevent or foresee excessive pressures on the

analyst, once the treatment has started, for moving away from
a position of technical neutrality.
While most borderline patients, in my opinion, do best
with modified psychoanalytic psychotherapy along the lines I
have proposed for these conditions, there are borderline
patients who can be analyzed without modifications of
technique, and I have become more optimistic in this regard
in recent years. Perhaps the most important criteria of
whether a borderline patient is analyzable are, in addition to
the potential for early severe acting out and the extent to
which there is an observing ego, two considerations: First, the
extent to which there exists a certain superego integration, so
that the patient presents only very limited antisocial trends.
When antisocial trends are marked, there is usually a danger
that conscious distortion and lying will become an important
feature of the treatment: under these circumstances, psycho-
analysis proper becomes very difficult indeed, or impossible.
Second, the extent to which object relations have evolved so
that at least some advanced, neurotic transferences-in
contrast to more primitive ones- are available.
The implication is that, when the patient has some
capacity for differentiated relationships in depth with other
human beings, there is less risk of the disorganizing effects of
primitive transferences on the analytic situation. In simple
terms, when differentiated oedipal features are strongly
present from the beginning of treatment, and realistic,
integrated kinds of transferences are available in addition to
the chaotic, bizarre, and fragmented ones of borderline
conditions, a standard psychoanalysis may be indicated. This
is particularly the case in infantile personalities with hysterical
features functioning on a borderline level.
Care in diagnosis and equal care in deciding on the mode
of treatment to be used will help to resolve some of the
potential sources of complication in the treatment. Once the
decision has been made that a psychoanalytic psychotherapy,
 TREATMENT OF BORDERLINES 825

rather than psychoanalysis proper, will be carried out, the

next question is, to what extent will there be a need for
structuralization of the patient’s life outside the treatment
hours, for setting of limits and the introduction of auxiliary
therapists, hospital setting, or psychiatric social work as part
of the treatment arrangement. The issue is the extent to which
the analyst will be free to deal with the transference in a
technically neutral way in the hours (because the dangerous,
destructive or self-destructive potential of the patient will be
taken care of elsewhere). Many of these arrangements
represent significant parameters of technique, and the
analyst’s clear awareness of them and their strategic
consideration as situations he should eventually try to
interpret will also help to maintain neutrality in other areas.
The therapist’s deviation from technical neutrality in the
form of directive behavior may be of such intensity and/or
duration that technical neutrality cannot be achieved or
recovered and modifications-rather than parameters- have
been established. Under optimal circumstances, gradual
interpretation of temporary deviations from technical
neutrality are possible, leading to a reversion to a neutral
stance which maximizes the possibility for transference
interpretation and resolution of primitive transferences.
Paradoxically, it may be very helpful for the analyst who
treats borderline patients in l o n g term psychoanalytic
psychotherapy or in analysis to also have an active, general
psychiatric experience with brief, crisis-intervention type
treatments with other such patients. In other words, when the
analyst feels secure doing short-term, supportive psycho-
therapy or crisis intervention and can remain firmly in an
analytic stance the pressure on him for “action” decreases. A
broad level of experience, the capacity to carry out alternative
modalities of treatment, and a careful, complete diagnostic
evaluation, all contribute to permitting the analyst to
maintain a serene and firm position while interpreting the
patient’s acting out and complex transference developments.
826 OTTO F. KERNBERG

One rationalization for countertransference acting out is

the analyst’s impression that the patient is not able to handle a
certain situation analytically because it corresponds to a
particular “ego defect.” Frequently, such ego defects turn out
to be complicated transference resistances that were not fully
analyzed because, in part, of countertransference develop-
ments interfering with full analytic exploration at periods of
acting out. If the patient does present important ego defects,
these should be evident enough to make it possible to evaluate
them as part of the diagnostic process, and then to explore
them in the early stages of treatment as part of the evaluation
of the patient’s perceptions in the hours, particularly, his
perceptions of the analyst’s interpret ations.
At times, mistakes in indicating psychoanalysis or
psychoanalytic psychotherapy for borderline patients are
unavoidable, and there are cases which, after a period of time,
require a revision of the treatment modality. If a psycho-
analysis that seemed at first indicated later appears contra-
indicated, it is relatively easy or safe to shift the modality
into psychotherapy. The principal reasons for this shift may
be the analyst’s awareness that persistent severe acting out
cannot be controlled by interpretation alone and threatens
the continuity of the treatment or even the patient’s physi-
cal or psychological survival. At times, reality conditions
(often brought about in part or totally by the patient’s illness)
deteriorate, and create vicious circles that interfere with
analytic work. At other times, the prevalence of primitive
transferences and primitive defensive operations evolve into a
fullfledged transference psychosis, with a loss of reality testing
in the treatment hours and the impossibility of reconstituting
an observing ego on the basis of an interpretative approach.
Under these circumstances, the analyst may re-evaluate the
situation with the patient, clarify the nature of the problems
that unexpectedly changed the total therapeutic situation,
and transform the analysis into a psychotherapy. In practice,
this usually also means sitting the patient up and carrying out
 TREATMENT OF BORDERLINES 827

face-to-face interviews: I see no advantage and only disad-

vantages in carrying out a psychoanalytic psychotherapy on
the couch rather than face to face (1975b).
In the case of the opposite situation, that is, when what
started as a psychoanalytic psychotherapy should in the
therapist's opinion be transformed into a psychoanalysis, the
situation is more difficult. Usually, under these circumstances,
there has been a development of non-neutral stances on the
part of the analyst, and the establishment of parameters or
even modifications of technique which may distort the
transference relation to such an extent that analysis becomes
difficult or impossible. Therefore, the more technically
neutral the position of the psychotherapist in a psychoanalytic
psychotherapy with borderline patients, the easier will be such
a shift into psychoanalysis if indicated. Now the analyst has to
ask himself searchingly to what extent he has utilized
supportive techniques (in the form of manipulative or
suggestive comments) or made supportive use of the
transference rather than analyzing it. If the analyst can
answer these questions satisfactorily in the sense that the major
transference distortions or departures from neutrality can be
resolved analytically, the case may still be shifted into
psychoanalysis.
At other times, particularly if a shift from psychotherapy
to psychoanalysis seems indicated after months or even years
of treatment, that is, when the patient has sufficiently
improved in psychotherapy to make psychoanalysis possible, it
may be preferable to 'evaluate the situation fully with the
patient, set up certain goals that should be. accomplished in
psychotherapy itself before terminating it, and then consider
the possibility of a termination of psychotherapy with
indication of starting a psychoanalysis-if still needed-later
on. Under these conditions, it is ideal, if the patient remains
without treatment for a period-at least six to twelve
months-so that the total effect of the psychotherapy can be
re-evaluated after the patient has functioned independently
828 OTTO F. KERNBERC

for a time and the mourning processes connected with

termination of psychotherapy have a chance to resolve.
Ideally, under these circumstances, the patient should start
psychoanalysis with a different analyst.
The application of psychoanalytic theory and technique
to the diagnostic study and treatment of borderline conditions
has, it seems to me, opened new modalities of psychothera-
peutic treatment for these patients, and has extended the
indications of psychoanalytic treatment itself. I would hope
that some of the findings derived from applying a
psychoanalytic technique to borderline conditions have
applications in nonborderline conditions as well. There are
many situations in the psychoanalytic treatment of basically
nonborderline patients where primitive transferences and
defenses become apparent, and the understanding gained in
psychoanalysis and psychoanalytic psychotherapy of border-
line conditions should be helpful for these other situations as
well. A task still wide open is the formulation of an
integrative, comprehensive theory of psychotherapeutic tech-
nique on a psychoanalytic basis.

REFERENCES

Bibring, E. (1954), Psychoanalysisand the dynamic psychotherapies. ThisJournal,

2:745-770.
Brenner, C. (1969), Some comments on technical precepts in psychoanalysis.
ThisJournal, 17:333-352.
Eissler. K. R. (1953), The effect of the structure of the ego on psychoanalytic
technique. ThisJournal, 1:104-143.
Fenichel, 0. (1941), Problems of Psychoanal~tic Technique. Albany: Psycho-
analytic Quarterly Inc.
Freud, A. (1936), The Ego and the Mechanism of Defense. The JVn'tings of
Anna Treud, 2. New York: International Universities Press, 1966.
Freud, S. (1963), Ps~choanalysisand Faith, ed. E. L. Freud & H. hleng. New
York: Basic Books.
Frosch, J. (1971), Technique in regard to some specific ego defects in the treatment
of borderline patients. Psychoanal. Quart., 45:216-220.
Gill, hi. hl. (1954). Psychoanalysis and exploratory psychotherapy. Thtj Journal,
2:771-797.
Gioyacchini, P. L. (1975). Psychoanalysis of Character Disorders. New York:
Jason Aronson.
 TREATMENT OF BORDERLINES 829

Greenson, R. R. (1965)' The working alliance and the transference neurosis.

Psychoanal. Quart., 34:155-181.
(1970), The unique patient-therapist relationship in borderline patients.
Presented at the Annual hleeting of the American Psychiatric Assn.
(unpublished).
Heimann, P. (1956), Dynamics of transference interpretations. Internat. J.
Psycho-Anal., 37505-310.
Jacobson, E. (1971), Depression. New York: International Universities Press.
Kernberg, 0. F. (1968), General principles of treatment. In: Borderline Con-
ditions and Pathological Narcissism. New York: Jason Aronson. 1975, pp. 69-
109.
(1975a), Borderline Conditions and Pathological Narcissism. New York
Jason Aronson.
(1975b), Overall structuring and beginning phase of treatment of border-
line patients. In: Borderline Conditions and Pathological Narcissum. New
York Jason Aronson, 1975, pp. 185-211.
(1975c), Transference and countertransference in the treatment of border-
line patients. Strecker hlonograph Series No. XZI of the Institute of Penn-
sylvania Hospital. (Reprinted in: J. Natl. A m . Pn-v. Psychiat. Hosp.,
7:14-24.)
-Burnstein, E., Coyne, L., Appelbaum, A., Horwitz, L., & Voth, H.
(1972). Psychotherapy and psychoanalysis: Final report of the Menninger
Foundation's psychotherapy research project. Bull. Menninger Clinic,
36:l-275.
Knight, R. P. (1953a). Borderline states. In: Psychoanaljtic Psychiatry and
Psychology, ed. R. P. Knight & C. R. Friedman. New York: International
Universities Press, 1954, pp. 97-109.
(1953b), Management and psychotherapy of the borderline schizophrenic
patient. In: PsychoanaZytic Psychiatry and Psychology, ed. R. P. Knight &
C. R. Friedman. New York: International Universities Press, pp. 110-122.
Little, hl. (1958), On delusional transference (transference psychosis). Internat.
J. Psycho-Anal., 39:134-138.
(1960), On basic unity. Internat. J. Psyho-Anal., 41377-384, 637.
Segal, H. (1967), Melanie Klein's technique. In: Psychoanal~ticTechniques:
A Handbook for the Practicing Psychoanat'yst, ed. B. Wolrnan. New YorW
London: Basic Books, pp. 168-190.
Stone, L. (1954), The widening scope of indications for psychoanalysis. This
Journal, 2:567-594.
Wallerstein, R. (1969), Introduction to panel: Psychoanalysis and psychotherapy.
Internat. J. Psyho-Anal., 50:117-126.
Wnnicott, D. W. (1958), Collected Papers. New York: Basic Books.
Zetzel, R. R. (1956), The concept of transference. In: The CapacityforEmotional
__
Growth. New York International Universities Press, 1970, pp. 168-181.
(1971). A developmental approach to the borderline patient. Amer. J.
Psychiat., 127:867-871.

Submitted January 23, 1976

New York Hospital-Come11hfedical Center, Westchester Division
21 Bloomingdale Road
White Plains, N. Y. 10605