THEORIES OF THE TREATMENT OF NARCISSISTIC PERSONALITIES

 Vann Spruiell, M. D.

Freud apparently intended to develop and expand his remarkable theoretical conception of Narcissism after World War I ended, but he never did. He probably also intended to formally expand his conception Das Ich which included his conception of "self," and, of course, English speaking analysts called the "ego," in more powerful terms in relationship to the dynamic unconscious -- which he named Das Es. That word was, unfortunately, translated into English as the "Id." Freud was more involved in his latter years with social and cultural issues. After the second World War ended, a number of analysts attempted to fill the theoretical gaps, not only of Narcissism but in English a 'Self" that many American analysts thought of as separate from the Ego. There was also a strong development conceptions of object relations which were also seen as also separate from the older hegemony of drive psychology. At first, these ideas were put forward primarily by Melanie Klein and her followers, but later influenced a variety of others, especially those of Otto Kernberg and Heinz Kohut. By 1970 these had become rival views in the United States.

One result was the presentation of a panel, "Technique and Prognosis in the Treatment of Narcissism" at the Fall Meeting of the American Psychoanalytic Association in New York, December, 1972. Originally, it was planned to be a sort of debate between Kohut and Kernberg, but Kohut elected to be represented by one of his most talented followers, Arnold Goldberg, (1974). Kernberg presented his own work. I was asked by the Chairman of the Panel, Martin Wangh, to present my own views, which were independent of those put forward by Goldberg and Kernberg.  

Another result (on my part) was the present paper. In it, I arbitrarily chose 12 issues by which to compare and contrast the clinical approaches to "narcissistic personalities" of Kohut and Kernberg, and their respective co-workers. Allusions to alternative psychoanalytical approaches were made, particularly in terms of the role of the Oedipus complex and the theoretical difficulties created by attempts to conceptually separate the ego from the self. Incidentally, the panel also was the occasion of a switch in formulations by Kohut through Goldberg away from the point of view of narcissism toward their description in terms of a particular form of object relations. Although my paper did not win me friends among the followers of either Kohut or Goldberg, I still support the ideas in it and the cogency of its criticisms and other observations. And work on it led to other papers which expand the ideas in this one, particularly: Three Strands of Narcissism (1975a), Narcissistic transformations in adolescence (1975b), Idealization (1978), Object relations theory (1979c), The self and the ego, (1981a), The indivisibility of Freudian object relations and drive theories, (1988), and Self (1991b).

 

In recent years there have been a number of approaches to the problem of narcissism, theoretically and clinically. Two of them were presented at the panel, "Technique and Prognosis in the Treatment of Narcissism"1. Assuming that neither Kohut's nor Kernberg's points of view have a corner on the market of truth, their agreements and sharp disagreements are healthy signs. Clearly, as one whose viewpoints are significantly different from either, I cannot even begin to do justice to the work which underlies the presentations, to either the rich and detailed clinical observations or the attempts at systematic explanations. All I can do in a preliminary way is express admiration for the former and reservations about the latter.

Systematic explanations can shadow as well as illuminate. For this reason, I hope to tease out, not systems, but some specific ideas -- 12 arbitrarily chosen issues that relate to diagnosis, basic conceptions, and therapeutic techniques in the psychoanalysis of so-called narcissistic personalities. These ideas on issues derive from a consideration of the two previous presentations in terms of: their agreements, their disagreements, and their negative agreements -- in the sense of what they both leave out. Finally, I will comment briefly about two of these "negative agreements."

Description and Diagnosis

Both Kohut and Kernberg describe nonpsychotic patients who have as their major problems specific peculiarities in the ways they attempt to maintain their self-regard. These patients not only have associated deficiencies or inabilities in loving other people, they demonstrate some major degree of superego pathology. There may be few surface symptoms, or the symptoms may mimic any of the standard psychoneuroses. These patients are called "narcissistic personalities."

Considered as a syndrome, the narcissistic personalities discussed by Kernberg make up a smaller group than do those patients describe by Kohut. The larger group would, it seems, encompass the smaller. Even if this is so, the same patients would be described differently.

Kernberg discusses individuals said to be different from other patients who regress to the fixation points of "normal infantile narcissism" -- although what the latter would be called is left unspecified. At any rate, the narcissistic personalities are thought to have a specific pathological sort of infantile narcissism. In differentiating patients with this sort of organization, close attention is paid to diagnostic considerations and to prognostic indicators from a variety of viewpoints.

The more general view, that of Kohut and his co-workers, obviously includes a much larger group of patients. In their approach to these patients, they do not rely on particular collections of signs and symptoms. Instead, the diagnosis is made when specific resistances in the transference neurosis allow the discernment of one of two transference (or transference-like) responses. In them, the analyst is treated, not as an independent, autonomous object, but rather as an aspect, so to speak, of the patient himself.

But diagnosis remains a problem for most psychoanalysts. Regrettably, neither using standard clinical techniques by themselves, nor using the transference neurosis by itself, has proven very successful in the past. One only has to think of the multitude of character syndromes or the "pseudoneurotic schizophrenic" or the "as-if personality" -- diagnostic labels which once aroused enthusiasm -- to maintain a sense of caution.

Historically, it seems that diagnoses made along one or a few dimensions are inadequate. Perhaps our whole approach to diagnosis is shortsighted. As an exception, Anna Freud's (1965) concept of developmental lines -- and thereby multidimensional diagnostic "profiles" -- has already borne fruit with children; Brenner (1972) has recently bought forward arguments in favor of modified similar approaches being used with adults.

The first issue, then, has to do with the tools, techniques and conceptions to be used in diagnosis.

The Clinical Understanding of Narcissistic Personalities

Kohut describes narcissistic analysands who demonstrate one of two -- only sometimes both -- forms of narcissistic transference, the "idealizing" and the "mirror," disguised by often very difficult resistances: coldness, haughtiness, inordinate self-sufficiency, sensitivity, contempt, etc. In the case of idealization, the analyst is seen as though he were a missing part of the patient's own psyche, a part that normally would have belonged to the idealized parents (thus, narcissistic in nature), and which would have later, as a result of optimal, phase-appropriate frustrations, been internalized as the normal idealization of superego functioning. In the case of mirroring, the analyst is seen only in terms of the responsiveness to, or his similarity to, or his actual congruence with, a warded-off "self," privately marvellous and sublime and admirable. The analyst corresponds to "the gleam in the mother's eye" as she interacts with her child's exhibitionistic display. Normally the "grandiose self" is gradually "tamed by virtue of optional, phase-appropriate frustrations. In the patient, however, it has been warded off by "horizontal splitting" (repression) and/or by "vertical splitting" (a concept roughly comparable to Freud's notion of ego splitting).

It is important here to note that the narcissism is thought to be transformed, either normally or in the working of the narcissistic transference neurosis, rather than given up. It is also important to note that, in Goldberg's terms (1974), "narcissism and object love have side-by-side existences and development" (p. 245). Yet, it is emphasized that they are separate. Again, quoting Goldberg, "narcissism has a development of its own, a pathology of its own, and requires a treatment of its own" (p. 245) Finally, it is important to note that while there are various statements referring to the dual instinct theory, the role of aggressive derivatives is usually described in reactive terms (Kohut, 1971a, 1971b).

Previously, (Kohut, 1971a, 1971b), this separate development of narcissism had been described in terms of a separate kind of libido, differentiated from object libido, apparently not convertible. Each kind of libido has many complex qualities.2 The "use of the terms object-instinctual investment and narcissistic libido does not refer to the target of the instinctual investment; they are abstractions referring to the psychological meaning of the essential experience" (Kohut, 1971a, p. 39n). Goldberg, on the other hand, did not use qualitative psychoeconomic explanations; instead he referred to the "narcissistic object" and its vicissitudes -- apparently intending the same thing Freud meant in 1914 by object choice.

Turning to Kernberg's theory, its use of "narcissistic personality" refers to a much more pathological infantile state. This state is first manifested in the analysis by resistance which functions "to deny the existence of the analyst as an independent, autonomous human being. This denial of dependence on the analyst does not represent an absence of internalized object relations or an absence of the capacity to invest in objects, but a rigid defense against more primitive, pathological object relations" (1974, p. 259).

Characteristic of the resistances "is the alternation of idealization of the analyst and self-idealization of the patient reflecting the activation of a pathological, grandiose self." This conceptualization is similar to one put forward in a justly well-known paper by Annie Reich (1960), although she did not base a syndrome on it. Her patients demonstrated a wide range of symptomatology and degree of pathology, and her clinical examples referred to phallic grandiosity and idealization.

The term "grandiose self" was not used in previous papers by Kernberg, but it is not the same kind of "grandiose self" Kohut has in mind. To Kernberg, this "self" is pathological by virtue of its being a condensation of some aspects of the primitive "real self," the "ideal self," and the "ideal object."

It should be interjected here that, suggestive as such formulations might be, they must be approached with some reserve. There are two reasons for this, one theoretical, the other clinical. In terms of theory, obviously we need concepts of self representations ("real" and "ideal") just as we need concepts of object representations. Schafer (1968) and Sandler and Rosenblatt (1962) have taught us a great deal about the representational world. Both make it very clear, however, that they do not mean such gestalten as causative agent, as little "systems." The invaluable conception of the representational world is to the psychic structure, in particular to the ego, as a map is to a terrain.3

Clinically, it is easy to reify concepts like "real" or "ideal" self and object (Glover, 1966). Actually, we have analytic data from adults and verbal children -- associations, fantasies, dreams, etc. -- to which we can usefully apply such terms. But we know very little about their structure or operational use in young children or in regressed adults.

To return to Kernberg's assumptions about a condensation of these elements, something equivalent to an infantile neurosis is thought to be jelled -- though, since it is not organized into the oedipal phase, it is not exactly equivalent. This "structure," pathological though it may be, permits, it is thought, a certain integration of the ego which borderline patients do not have. However, the underlying organization is considered to be the same in both conditions; the use of ego splitting (a term referring more to the splitting of objects and drives than to splitting of ego organization in Freud's sense), projective identification. omnipotence, primitive idealization, and defenses relating to envy.

Aggression is considered by Kernberg to be of paramount importance, though it is clearly stated that the analysis must relate to both libidinal and aggressive derivatives. The theory also emphasizes that the patients do not suffer from an absence of certain structures in the ego and superego, but the presence of pathological primitive structures. Finally, the theory assumes no separate development of the representations of self and objects; it assumes that one cannot study the vicissitudes of narcissism without studying the vicissitudes of object relationships and vice versa.

It will be noted that the two theories use some apparently similar concepts: vertical splits in relation to ego splitting, mirroring in relation to projective identification, idealization in one theory or the other. One must be cautious, however; the actual definitions are different. Their basic metapsychological contexts tempt one to think they derive from quite different views of man. The divergences in the two theories should be obvious. They raise the following major issues:

the basic nature of narcissism -- the way or ways it is to be considered metapsychologically;

whether narcissism and object love/hate have separate (though related) existences and developments, from the viewpoint of the drives or whether narcissistic and object drives are seen as convertible and as less separable;

whether narcissism develops in identifiably separate "idealizing" and "grandiose-self" forms:

whether aggression is seen primarily in instinctual or reactive forms:

how to conceptualize the nature of the early representational world -- of representations of primitive "real" and "ideal" selves and objects;

whether pathological narcissism in the adult is a consequence of some sort of pathological infantile narcissistic organization (like the equivalent of an infantile neurosis), or whether it is merely a result of interference with the evolution of normal infantile narcissism -- or whether there is some sort of combination of these possibilities.

There is "negative agreement" about two other issues: the use of a concept of "self "beyond the meaning of self representation and aside from the concept ego, and, the relative unimportance of the Oedipus complex in these conceptions of pathological narcissism.

At this point it is well to recall that there are still other metapsychological approaches which differ, in some respects profoundly, from those we have been considering. Accordingly, the same patients might be understood quite differently by these other approaches.

Treatment

Both theories share the assumption that the problems in the narcissistic personality did not arise in that remarkable coming together of comedy, awfulness, exaltation, depravity, tragedy, envy, lust, murder, expiation, grandness, subjugated despair, and fantastic innovation we imply when we refer to the Oedipus complex. Kohut indeed assumes the oedipal period to be the major time when massive traumatic frustrations may occur. But in the patients he describes, he considers these to be narcissistic problems, not problems having to do with loving or hating others as separate entities. Presumably the relations with objects may or may not be affected. Most of the time, this is seen as an either-or proposition, inasmuch as narcissism is though to have such a large measure of independence.

Kernberg locates the primary etiology very early. The underlying conflicts are around oral rage and envy in connection with severely pathological internalized object relations as a preoedipal, particularly oral level.

Naturally, the therapeutic approaches derived from these assumptions will be very different. If on the one hand object love/hate relations, including the transference relationships, are assumed to exist behind their complex disguises on an oral sadistic level, as in Kernberg's theory, then it will be expectable that these patients will be extraordinarily slow to respond to interpretive interventions. It will be expectable that they will be destructively hostile to the analysis and the analyst. "All the patient's effort seems to go into defeating the analyst, into making analysis a meaningless game, into systematically destroying whatever they experience as good and valuable in the analyst" (Kernberg, 1970, p. 70). The patient is too consumed with envy to tolerate true dependence, although he may seem dependent. He may seem to idealize the analyst, but these are pseudo-idealizations aimed only at extraction. They are basically negative, hostile. The analyst must consistently -- one might almost say implacably -- ferret out and interpret the negative transference. He must also counteract the patient's efforts toward omnipotent control and devaluation. This approach to the negative transference was particularly emphasized in earlier papers to the extent that one wondered how an analysis could take place under those conditions. The paper under consideration modifies the emphasis considerably.

To continue Kernberg's view: if the work is successful, it eventually results in the outbreak of dangerous, regressive, paranoid rage. If this can be dealt with, it is followed by intense depression and guilt, with suicide becoming a distinct possibility. This represents a crucial point in the therapy when there is "merging" of the loved and hated "internal object." If the patient is capable of working such material through, he presumably becomes able "to acknowledge the analyst as an independent being to whom he can feel love and gratitude" (1970, p. 81).

In the panel paper (1974), not previously, Kernberg emphasized that the basic goal of the work is to reach the point where the transference may shift into the ordinary transference paradigms characteristic of transference neurosis.

One is doubtful that the patients working in analysis in the ways just described would be diagnosed as narcissistic personalities by Kohut and his fellow workers in Chicago. However, assuming for the moment they would be, the analytic understanding of the patient's coldness, his arrogance, his rage, his maneuvers to deny the validity of the analysis, would be very different from the understandings of Kernberg -- just as the understanding of the later idealizations or treatments of the analyst as a mirroring object would be different. The understanding would be that the earlier reactions defend against the therapeutic reactivation of incomplete and split-off aspects of infantile narcissism; if the resistances are analyzed, these emerge in the idealizing or mirroring forms of the transference -- essentially different from the transference produced by ordinary neurotics.  

The technical approach is traditional; the emphasis of the work is, at least in form, exactly along the lines of classical technique: acceptance of the resistance or defense, the correct understanding of it, the avoidance of premature or intellectualized interpretations, the understanding and mastery of one's own countertransference, the allowing of the transference neurosis to spontaneously develop, the gradual maneuvers toward the end to allow dissolution.

According to Kohut and his co-workers, there may be a coexistence of narcissistic and "object-transference" pathology, and manifestly narcissistic phenomena may disguise a nuclear oedipal conflict. The almost grudging tenor of these considerations may be conveyed in the following quote: "It must . . . be mentioned , however, that even in some cases of genuine, primary narcissistic fixation, an oedipal symptom cluster (e.g., a phobia) may still emerge, if ever so briefly, at the very end of the treatment and must then be dealt with analytically as in the case of a typical primary transference neurosis), Kohut, 1971a, p. 155).

Some of the clinical divergences in the two theories are as follows:

the basic understanding of the transference and resistances,

the timing and content of interpretations, and particularly,

the understanding of idealization as defensive or as part of progressive maturational movement.

I began with the statement that both theories agreed that the pathology was not organized at the oedipal level -- at least in terms of customary understandings of that phase. A third possibility is that it is organized at that time. In this view, regressive and other phenomena disguise the patient's innermost conviction that he has solved oedipal dilemmas by triumphing over his rival or rivals. Among analyzable individuals, at least, a narcissist might be defined according to this view as one who has to reorganize his life around the belief that he has won an oedipal victory.

Discussion 

It seems to me that the understanding of narcissism has been greatly furthered by Kohut's and Kernberg's work, despite, or perhaps because of the glaring disagreements. But it also seems to me that very valuable clinical observations can be damaged by high-level metaphysical assumptions that "shadow as well as illuminate."

The understanding of narcissism will be furthered still more if we recognize that there are a variety of techniques me use to try to maintain their self-regard: there are thus a variety of pathological possibilities. It is clear by now that psycho-economic explanations alone are not sufficient to conceptualize the problems, and that a more adequate theory of affects is sourly needed (Joffe & Sandler, 1967; Brenner, 1972). New conceptions in metapsychology are developing, and may go hand in hand with promising clinical developments -- examples are the work of DeSaussure (1971) and those we have heard today.

I will conclude with remarks about two of the issues already raised, the role of the Oedipus complex and the concept of "self." My clinical experience convinces me that, despite the undoubted importance of pre-oedipal conflicts, the Oedipus complex is still to be considered central in the psychic organizations of analyzable patients, woven when -- or especially when -- it is not visible or cannot be effectively interpreted until late in the treatment process. The work of Tartakof (1966), Annie Reich (1960) and Arlow and Brenner (1964) would seem to corroborate this view.

As for the related issue of "self," Goldberg (1974) claims that " the overriding importance of the self as a content of the ego closer to experience . . . might strain the applicability of structural theory to the clinical phenomena" (p. 14). Might it? Must it?

In the last 50 years we have witnessed the steady erosion of the concept of id by concepts of ego, the "dissolution" of concepts of superego, the denigration of the dynamic unconscious - whether seen as id or ego -- by other aspects of ego. Now the ego is itself threatened by concepts of "self." Did all this happen because the structural model is so inadequate? Or did it happen because we are reluctant to accept its full implications? One of these, of course, has to do with the fate of the Oedipus complex.

After all, the ego was the self. Later, defined more abstractly in terms of its functions it came to mean more -- but at the same time less -- to those for whom it no longer represented the self. But do not most of those concepts of functions still represent the functional underpinnings of the self in relation to its own need, to its, drives, its objects, its relativity, and its moral imperative?

ENDNOTES

1. Heinz Kohut, represented by Arnold Goldberg, M.D. and Otto Kernberg, M.D. were the primary speakers.

2. (1998). These postulated "qualities" bore little resemblance to the conceptions of "qualities" by conceived by contemporary neural scientists to explain the complex connections and interactions of neurons.

3. (1998). Metaphors can be misleading. No contemporary neural scientists, to my knowledge, believe that external objects are literally reproduced in the brain. In fact, the productions of those codes created to in some way represent the external world, and their eventual translations into representations of the external world that can be perceived as phenomenal events remains one of the great mysteries. The mind-brain relationships still do not seem close to adequate explanations even now.


REFERENCES

Arlow, J. A. & Brenner, C. (1964), Psychoanalytic concepts and the Structural Theory. New York: International Universities Press.

Brenner, C. (1972), Some observations on depression, on nosology, on affects and on mourning, Unpublished.

DeSaussure, J. (1971), Some complications in self-esteem regulation caused by using an archaic image of the self as an ideal. Internat. J. Psycho-Anal., 52:87-98.

Freud, A. (1965), Normality and Pathology in Childhood. New York: International Universities Press.

Freud, S. (1914), On narcissism. Standard Edition, 14:67-102. London: Hogarth Press, 1961.

Glover, E. (1966), Metapsychology or metaphysics? Psychoanal. Quart., 35:173-190.

Goldberg, A. I. (1974), On the prognosis and treatment of narcissism. This Journal, 22:243-254.

Joffe, W. G. & Sandler, J. (1967), Some conceptual problems involved in the consideration of disorders
of narcissism. J. child Psychother., 1:56-66.

Kernberg, O. (1970), Factors in the Psychoanalytic treatment of narcissistic personalities. This Journal, 18:51-85.

------------------ (1974), Contrasting viewpoints regarding the nature and psychoanalytic treatment of narcissistic personalities: a preliminary communication. This Journal, 22:255-267.

Kohut, H. (1971a), The Analysis of the Self. New York: International Universities Press.

------------- (1971b), Thoughts on narcissism and narcissistic rage. Presented as the Brill Lecture, New York Psychoanalytic Society, November 30th, 1971.

Reich, A. (1960), Pathologic forms of self-esteem regulation. In Psychoanalytic contributions.
New York: International Universities Press, pp. 288-311.

Sandler, J. & Rosenblatt, B. (1962), The concept of the representational world. The Psychoanalytic
Study of the Child, 17:128-145. New York: International Universities Press.

Schafer, R. (1968), Aspects of Internalization. New York: International Universities Press. 

Tartakoff, H. H. (1966), The normal personality in our culture and the Nobel Prize complex. In: Psychoanalysis - A General Psychology, eds. R. M. Loewenstein, L. M Newman, M. Schur, & A. J. Solint. New York: International Universities Press.