This is a paper written for the book, Psychoanalysis: The Major Concepts, edited by Burness Moore, M.D. and Bernard Fine, M.D., published in 1995 by the Yale University Press. I was a member of the Editorial Board of Moore-Fine project. I had written a previous paper on the subject, one that expressed a point of view of my own (although some distinguished colleagues shared its central ideas). The earlier paper was titled The Self and the Ego. vs81a. I was asked to write a more broadly based paper for Psychoanalysis: the Major Concepts. The subject was a sensitive one for various reasons. In the United States, there were controversial discussions about the contributions of Kohut, Kernberg, and the countering responses of more mainstream analysts. Some of the members of our Board were unprepared for a part of a central point that I made in both papers: that, much as Heinz Hartmann's contributions, especially in terms of his furthering adaptation as a central concept in development, I found much to disagree with in his formulations about the subjects of narcissism and the nature of the "self" and ego.

 I did try to present the arguments of those who supported Kohut's and Kernberg's points of view (and still others who proposed that the "self" should be a "superordinate" point of view, over the abstractions of ego, id, and superego). I invited my fellows to read my draft, and if they felt that there were arguments opposing mine that had been left out, to let me know and I would include them. Or, if there were still problems, I would withdraw my paper. I did receive responses, and, in fact, modified parts of my earlier draft. But the central points remain as I had first drafted them.

The paper is tedious, and I am sorry about that. It had to be so condensed because of space that there was no alternative but to write tightly. The earlier, simpler paper, is easier to read. Nevertheless, I think that the arguments presented are important enough to be examined carefully by anyone interested in psychoanalytic theory. Although it is facile for some analysts to mock "theory," it should not be mocked; rather, it should be examined and criticized. Our most fundamental assumptions and basic concepts ought to be closely related to clinical findings -- if they can be. They should not separated from our collective clinical experiences (see my Freud lecture, vs90). These basic assumptions and concepts represent the "infrastructure" of psychoanalysis almost as much as does our collective experience with patients. In my opinion, that infrastructure has been invaded by a lot of careless thinking -- and I am far from being alone in this opinion.  

INTRODUCTION

"Self" is a necessary word, used every day, a noun related to the pronouns, employed most often as an adjective and as a prefix. It cannot be avoided in ordinary discourse, not even by psychoanalysts or philosophers. And in ordinary discourse it is readily understood. But there is no consensus among English-speaking analysts about whether it should be used as a technical, theoretical, or abstract word -- or if it should be so used, how it can be defined in other than arbitrary ways.

Formal "objective" concepts of "self" are problematic in the general sense because of the reflexive nature of the word: it is almost impossible to define as a noun except in terms of itself -- tautologically. It has posed special problems for psychoanalysis, and continues to. In this paper I will survey these problems but I cannot hope to solve them to general satisfaction.

Freud originally used several different words which would approximately mean what we call "self" in its everyday sense: an individual mind capable of conceiving its own operations either introspectively or inferentially. He also used different abstract terms for the underlying psychic structures and functions which could account for the self in experience and in action. To subsume the varying words, he eventually settled on das Ich, which in German can stand both for the first person pronoun and as an abstraction on higher levels. Das Ich is commonly and comfortably used in both senses; the German word selbst, in contrast, is ungainly, and not used very often in everyday discourse. In many other languages, single words can convey a range of meanings from the phenomenally "obvious" to the most abstract levels, i.e. the French word for das Ich is le Moi; in Spanish it is e1 Yo. But in translation into English, das Ich became not "the I," or "the me," but the ego.

The decision to translate the word as ego was a monumental one. Ego is not a common word in ordinary English. Except among analysts it refers to vanity or narcissistic self-aggrandizement. The first person pronouns, and to a lesser extent, "self, are usual and ordinary words; but in English it seems strange to speak of "the I" or "the me." Hence, there is a tendency among English-speaking analysts to separate the terms: to limit the ego to abstract systemic or structural meanings, and to use self and self-representation in place of the phenomenal Ich. Going further, some analysts have extended the ordinary use of "self" to the use of it also as a highly abstract noun, referring to something distinguishable from the abstract ego. Metapsychology, at least as put forward in English by these latter-day analysts, has acquired an embarrassment of riches: two abstract systemic concepts, a reasonably well-defined abstract ego, and an equally abstract -- or even "superordinately" abstract -- but poorly defined self.

The various points of view may be summarized:

A group of psychoanalysts, otherwise similar in thought and practice, believe it is heuristically and logically valuable to separate self from ego. Clinical phenomena, they say, should be clearly separate from abstract theories; the ego is an abstraction referring to a set of psychic functions which cohere and (at least during periods of conflict with other systems) form a cohesive mental structure; the word ego implies systemic meanings. However, this group is not of one mind about the meaning of "self." Although all acknowledge its everyday experiential meanings, there is no consensus about whether it should have a place in psychoanalytic theory or if it were used theoretically, how it would be defined.

A minority disagrees, claiming that ego should be used as Freud used das Ich. They believe that Das Ich should have been translated as "the I" (a translocation of meaning somewhat easier to imagine than "the me"). The ego (or the "I") would refer at one pole of meaning to an abstraction on a clinical level corresponding to self in its everyday sense -- and on the other pole of meaning to a coherent system of psychic functions. It is an advantage, not a disadvantage, they say, to have a unitary term, Das Ich: the reason that the functions defining the abstract ego cohere is the very fact that they have to do with the operations that people intuitively think of as referring to "self." However, they recognize that long-standing customs do not change easily; analysts, like other people, will continue to use the words to which they are accustomed and English-speaking analysts will use the separate words, ego and self. The aim of the minority group of analysts is merely to call attention to the heuristic advantage of regarding the phenomenal I as absolutely related to the abstract I.

More dissident psychoanalysts believe that self should not only be distinguished from ego, but that on the highest levels of abstraction it can be delineated as separate from ego: it might be a fourth microstructure of the mind, joining and relating to the id, ego and superego. Alternatively, an abstract self might be on a still higher level, superordinate to ego, superego, and id.

I will not pursue semiotic and semantic issues further, but will turn to the history of the uses of ego and self in psychoanalysis. It will begin with Freud's explicit and implicit conceptions, then take up amendments proposed by Hartmann and others in the elaboration of ego psychology. From there, I will consider the positions listed above and will then examine the arguments favoring or opposing each.

FREUD'S VIEWS

How perplexing Freud would have found it to learn that he had "neglected" the self (Levin, 1963), or that he "never developed an elaborate theory of the self" (Mitterauer and Pritz, 1978)! On the contrary, Freud spent his entire working life spawning more and more elegant theories about those meanings that in English subsume self or the first person pronouns. For most of the nineteenth century, psychology was a psychology of the conscious. The self was what one privately thought it to be -- on the conscious level. There were thinkers before Freud, of course, who thought of unconscious aspects of mind (Ellenberger, 1970). Nietzsche, to take a late example, conceived a truly dynamic unconscious, but had no way to validate his philosophical speculations. Freud, by contrast, was able to find validations. He patiently explored conflicts within the mind among forces and counter-forces. On the one side was the self that could be experienced (at least potentially): what he variously called will, volition, consciousness, and "das Ich" (early on he used quotation marks in an effort to avoid philosophical conundrums connected with German philosophical usage of Das Ich). On the other side of such conflicts were the psychological influences variously referred to as "counterwill," "antithetical ideas," or "the unconscious." Thus, he constructed the first metapsychology (Brenner, 1980). His monumental, though premature, effort to construct a neuropsychology, the "Project for a scientific psychology (1895), surely postulated a complex system to explain the self or ego. Why else would he call that system das Ich?

Although analysts have become accustomed to thinking of the system, id-ego-superego as the structural theory, it was only Freud's last structural theory. He (and Anna Freud) saw the topographical system abstractly as well as phenomenologically -- and continued to use the earlier structural system side by side with the later id-ego-superego, bringing out the one or the other depending upon the clinical contexts. It is unnecessary here to trace the development of Freud's topographical structural theory or the later disagreements with Freud about its continued value (see Arlow and Brenner, 1964; Gill, 1963). The point is that the Cs-Pcs system was an abstraction referring to the known self. And the Ucs., the dynamic unconscious -- at least the part that was once conscious -- referred to the unknown (but potentially partially knowable) self.

It is also unnecessary here to trace the development of the ego concept in Freud's work (see A. Freud, 1936, 1952; Hartmann, 1950, 1956; Laplanche and Pontalis, 1973). After developing the Cs-Pcs. and Ucs. systems, Freud used das Ich (the "I") casually, in an everyday sense. But when he turned away from the topographical systems (without, as mentioned above, abandoning them), he proposed elaborate abstract conceptions of the ego as a coherent organization (1923, 1926a, etc.). He continued thereafter to use ego-as-self and ego-as-system, relying on the context to define the level of abstraction.

Some authors have suggested that Freud was not aware of the distinctions.

What a poor opinion of Freud! He knew what he was doing. Referring to the ego in 1926(b), he said, 

For all these firm words, Freud had agreed in 1919, for whatever reasons, to Strachey's translation of das Ich into the Latin "ego." Strachey (1961) admitted to difficulty with this translation; sometimes, he translated das Ich directly as "self." Presumably, Freud also agreed to the Latin "id" as translation of das Es. "The It" would have been a much better choice. Recently, sharp criticisms have been put forward about these renderings, claiming that they do not do justice to Freud's full meanings of the terms and, in particular, disguise their affective meanings (Spruiell, 1981; Ornston, 1982; Bettleheim, 1982.

SEPARATE DEFINITIONS FOR SELF AND EGO

Hartmann in 1939 (1958), later in collaboration with Kris and Loewenstein, set out to systematize metapsychology as a general psychology. As one part of a life's work that was monumental in reach, Hartmann was interested in certain economic assumptions having to do with the more or less neutralized energic processes of the ego as an organization, and those other, less neutralized energies having to do with narcissism. Following Loewenstein (1940), he pointed out that it was important to differentiate, as Freud had not, the concepts of ego, self, and person. His arguments will be discussed more specifically below.

The matter of narcissism does raise puzzles. Observably, a person can take himself as his own love object. This is normal during infancy, and in altered forms it exists normally in later life. Some people, however, continue to love themselves in primarily infantile ways, and all people do so some of the time. But if this is true, an economic question is posed. How to distinguish between those psychic energies which "power" the functions of the ego, and which by their nature can only be experienced indirectly, if at all, from those derivatives of libidinal investments of the self which can be experienced and thus clinically identified as narcissism?

As Hartmann (1950) put it,

Here, Hartmann identified self as "one's own person," and also distinguished this "self" from its mental representations, just as external objects are distinguished from their representations. But in 1953, Hartmann spoke of "the cathexis of the self-image (a complex of representations)" (p. 185). The self then would be "a complex of representations," capable of being cathected with energy, and thus presumably capable of operating as an intrapsychic agent. In 1955 he wrote, "If we accept (this definition of narcissism as the cathexis of self), we may then speak of self-representation (in the case of libidinal cathexis: narcissism) in opposition to object representation (p. 21).

He made similar propositions again in 1956. Narcissism cannot be simply a matter of the ego, Hartmann averred, if the ego is viewed abstractly as a set of functions. It is a matter of the whole body and mind of the person -- body, ego, id, and superego alike.

SOME CONSEQUENCES OF HARTMANN'S CONCEPTS

The influence of Hartmann and his colleagues on the development of ego psychology, especially in America, can hardly be overestimated. His work, with Anna Freud's, allowed psychoanalysis to come closer to becoming a general psychology -- or becoming an important part of one. Without their efforts, it is hard to see what might have kept psychoanalysis from fragmenting into various irreconcilable forms of id psychology and object relations psychology. It is hard to see what might have kept psychoanalysis from fragmenting into various irreconcilable forms of id psychology and object relations theory. It is hard to see how mainstream psychoanalysis could have achieved sufficient coherence to allow it to take its place among related disciplines and gradually evolve and interact with them.

Nevertheless, new formulations may carry part-formulations that create unforeseen consequences. The separation of "self" from "ego" was such a part-formulation. In fact Hartmann's formulations have given rise to fateful consequences when used by most subsequent writers who have concerned themselves with self, not only those who adhere to the propositions of ego psychology developed after the 1940's, but by others, who, to various degrees, modified them. They influenced Edith Jacobson's (1964) work, which focused on the normal and pathological developmental distinctions between self and object representations -- and which tended to treat psychic representations as though they are small complex systems -- quasi-autonomous agents of the mind (Schafer, 1968). Jacobson's ideas have influenced the object relations theories which have grown up, among them, Kernberg's (1976). And they have had a role in the over-extensions by some writers of theories of separation-individuation (Mahler, et al, 1975; Spruiell, 1979a). They have had an influence on the blossoming of a "self psychology" by Kohut (1977) and his followers, who rejected Hartmann's drive psychology on the one hand, but vastly elaborated his notions of narcissism and self on the other.

The temptation to go even further has been manifest in proposals that the self should be created as the fourth metapsychological macrostructure (Levin, 1969; Mitterauer and Pritz, 1978). The more extreme of these outgrowths postulates a self that is a superordinate psychic organization, overarching the familiar ego, id and superego (Gedo, 1979; Kernberg, 1982; Klein, 1976; Kohut, 1977, 1978, 1982).

A more interesting version, in my opinion, of the attempt to distinguish self from ego was advanced by Spiegel (1959). He carefully dissected some of the semantic problems posed by the words, self, self-representation, and self-awareness. As Beres stressed later (l981), he began with the assumption that "self" is an abstraction to be distinguished from self-awareness. According to Spiegel, "Self-feeling is an ultimate, not further describable clinical fact, but the self is not a clinical fact in the same sense ... It is a conceptualization or a construct which we invoke to clarify clinical phenomena..." (p. 87). It has value both in the study of the differentiation, or lack of differentiation, between the self and its objects, and the study of alterations of self-awareness.

Along the same lines of thought, developmental studies, such as those of Mahler, Pine and Bergman (1975; Mahler and McDevitt, 1982), have provided a framework for understanding more about the emergence of the self and its distinction from objects. The concepts of separation-individuation assume that features of psychic functioning ordinarily related to the self constitute the framework and dimensions of self-fantasies. Spiegel (1959) emphasized the spatial and perceptual character of self-language, and its basis in fantasies about the body. Thus "self" is a "reference framework which has perspective" (1982).

Grossman (1982) elaborates Spiegel's sophisticated line of thought. "Self" is fantasy, like a theory, a collection of conscious and unconscious fantasies. It is constructed like other fantasies, but has a special importance in the regulation of behavior. Obviously, the stability of such fantasies, in interaction with fantasies of objects, and their interdigitation with reliable representations of the external world, are of prime importance in adaptation. It is important to reiterate that this concept of self, that of Beres, Spiegel and Grossman, is a clinical abstraction, like Freud's expeience-near version of das Ich; it does not amount merely to self awareness.

In summary, there is at least an agreement among these latter Freudian analysts that the self refers to some intrapsychic pattern or structure, one that presumably becomes more unified and stable if the winds of development are favorable. These analysts assume that the various conscious and unconscious representations (including misrepresentations) of the self, involved in sundry interactions with object representations, are related. It is this relatedness, and the putative synthesis which is supposed to occur -- this psychical pattern in a slow rate of change -- that is thought to constitute the identity of the person. For them, the self becomes an abstraction for experiential and potentially experiential phenomena. Unfortunately, they do not make it clear how this pattern in a slow rate of change is associated with their concepts of ego, as system.

CRITICISMS OF HARTMANN'S CONCEPTS

The "clarifying distinction" leading to separate definitions of ego and self confused some analysts. While most mainstream analytic critics agree that, to be adequate, an economic drive theory must distinguish between erotic and aggressive cathexes of the ego-as-self and those presumably more neutralized energies powering the ego-as-system, there are three distinct sets of objections to Hartmann's reformulations.

First, there is no particular reason to assume that what we think of as psychic energies -- for example, in the experience of some kind of self-love -- and energies that cannot be experienced but have to be inferred abstractly -- for example, those involved in internal conflicts -- are of the same sort. "Psychic energy" is a metaphor, just as "physical energy" is.

Second, there are objections to the equation of narcissism with the libidinal cathexis of the self (Spruiell, 1975, vs75).

Third, there are objections to each of Hartmann's definitions of the word self (Laplanche & Pontalis, 1973; Spruiell, 1981).

These theoretical issues can only be summarized briefly here. At the present time, the specifics of a psychological economic theory can be postulated only with the aid of distant metaphors. Nor are models of hydrodynamic or electrical systems the only possible metaphors for mental quantities and the part they play in psychic dynamics. The demands of what is still theoretical in psychic economics should not become the basis of a whole of theory.

Narcissism is still a bridge concept between what can be clinically experienced and observed and what can be thought of abstractly. Unfortunately, it has never been adequately integrated with the later theoretical concepts -- of which "On Narcissism" (Freud, l9l4) was the harbinger. That paper was written before the modern structural theory -- id-ego-superego -- was set in place, and before the later dual instinctual drive theory was presented. In fact, narcissism was originally contrasted with egoism and this distinction reappeared even after 1920.¹

Narcissism was indeed only "introduced." The concept, as noun or adjective, was applied to a normal developmental phase, a perversion, a type of object choice, the ego ideal, regulations of self-esteem, solely to omnipotence, or solely, as in Hartmann's formulations, to self-love. It would be more useful, and certainly less confusing, to think of all these as constituting related parts (Spruiell, 1975). But Hartmann's economic treatment of narcissism only in terms of self-love is a concretization, or at least a narrowing of the concept. He treats it as if it were self-evident and precise, easily transferable to the structural point of view. Yet, as an umbrella concept, it cannot account sufficiently for the myriad of clinical phenomena dubbed "narcissistic," as Joffe and Sandler (1967) documented clinically.

And what did Hartmann mean by the "self" when he spoke of it being cathected by libido? As we have seen, he variously meant self-representation, a collection of self-representations, and a person as a whole. But if self means person as a whole, then "self" is unnecessary, except as an indicative pronoun. If "psychic apparatus," then self is simply a word for some individual mind in its totality. It is true that Freud occasionally (1915, p. 134; 1921, p. 130) used Gesamt-Ich to refer to "the ego as a whole" -- distinguishable from the ego as one constituent of the systems id-ego-superego. However, as Loewald (1973) remarks, "If self is something like Freud's Gesamt-Ich ... then, far from being a content or a structure within the mind, self would be the mind as cathected in its totality" (p. 450). A totality cathecting a totality is problematic.

Did not Freud himself convince us how powerful are impersonal aspects of mind? Self-as-one's-whole-person (according to an outside observer) is an interpersonal, not an intrapsychic term.

Nor can a totality logically be equated with a part of the totality, a content, a psychic representation, or a collection of representations. The "representational world," as it is ordinarily conceived, is no substitute for the structural, or systemic, point of view. The concept of mental representation varies depending upon the metapsychological position taken (Friedman, 1980; J. Jacobson, 1983a, 1983b). A cogent paper by Boesky (1983) summarizes these variations, and the unfortunate clinical consequences that can result from thinking of representations as small systems of the mind. Such "systems" typically replace drive concepts. Self-representations can no more be the self than objects can be the same as their representations. The postulated "representational world" -- which might be better called part of the internal world, is analogous to an intrapsychic map, as Hartmann pointed out in 1939 (1958, p. 58). The map is hardly the terrain.

Even the representation of wholeness is questionable. Although a person regularly perceives himself as agent, he rarely experiences himself as a "whole." In states of altered consciousness, during orgasm or some forms of meditation, feelings of wholeness may be experienced -- but, paradoxically, those are the same times when the delimitations and details of the self become blurred. It is a misunderstanding of the concept of psychic representation to equate self with its varying representations (Loewald, 1973).

Those who define the self as the whole person are left with an interpersonal term with neither intrapsychic distinction nor theoretical place (Schafer, 1973). The question becomes: whole person according to whom? We cannot say, "according to the analysand," because one of the preoccupations of any analysis is to address the analysand's self-deceptions. We cannot say "according to the analyst," because he has immediate knowledge only of his own self, however valuable that knowledge might be for empathic understanding.

Why then speak of self at all? Because we can't help ourselves: Persons must identify themselves to each other and indicate who is being referred to and who is doing what, and what we think we know about ourselves or other selves. But usually when an analyst is speaking to an analysand about the latter's self, he means two things: he is addressing himself to the other's self awareness on the one hand, and on the other, he is implying that there is more to this "self" than the patient knows. At least in his own mind, the analyst is relating these experiences to a more complex set of guiding fantasies alone with their non-experiential operations, such as monitoring, stabilizing, resisting, compromising, acting upon, integrating, etc. In other words, the analyst, though not necessarily the patient, is thinking of what is signified by psychoanalytically useful abstractions. Beres, Spiegel and Grossman would prefer to use the word "self" for the experiential aspects; this writer would prefer to use "self" as a necessary, everyday word, but use the word "ego" to refer to both the phenomenal and the abstract ego, just as Freud preferred the range from the experiential to the non-experiential which das Ich made possible.

Hartmann (1953) also mentioned "the cathexis of the self-image (a complex of representations)" (p. 185). Other analysts, e.g. Kernberg (1982), lump the various self-representations together, and call that the self. Thus the self would be "embedded" as a "structure" in the ego. Presumably it is the ego-as-system, not the "structure," which would perform the synthesizing, the lumping. No individual can consciously do it for himself, of course, nor can it be done for him.

Finally, Hartmann, who argued persistently against tendencies to reify abstractions, spoke of our "finding narcissism in all systems." But how can narcissism, an abstraction, be "in" other abstractions? The difficulty with Hartmann's reformulation of ego and self is that it represents an overly systematized approach to what is, after all, a speculative mental economics. The result is a further detachment of the abstract concept of ego -- the non-experiential realm -- and the experiential ego (Sandler and Joffe, 1969). Although Schafer (1973) (who also would refrain from using abstract conceptions of self) asserts that Hartmann did not use self and identity as metapsychological notions, it is impossible to agree with him. How can the "self," whether as "whole person" or as a representation, be cathected without assuming for it some sort of metapsychological status?

Hartmann sought to resolve a theoretical ambiguity concerning the ego. But the effort not only brought forth new problems, it weakened the original concept. As Laplanche and Pontalis (1973) put it,  

The ill-defined or undefinable self can become an empty abstraction: a signifier confused with an unknown and unknowable something that it is supposed to signify. An imaginary "structure" of the mind that is an object, a location, an agent, an originator of action -- in short, a humunculus (Rangell, 1982). As for "self" as a theoretically highly abstract word, Glover, called it "a journalistic term" (1966).

THE EGO AS BOTH PHENOMENOLOGICAL AND ABSTRACT CONCEPT 

The arguments that follow do not dispute the assumptions about psychic function held by the some of the contributors cited above, especially those of Spiegel, (1959) and Grossman (1982). To repeat, the only disagreements have to do with whether there is heuristic value in separating "self" from "ego." Negative criticisms have been made of assumptions -- mostly among ego psychologists in America -- which distinctly separate the self from the ego. Positive arguments may be added for retaining Freud's dual meanings of das Ich (whichever word or words are used to translate it). However, it is unavoidable -- and ironic -- that these arguments have to be cast in personal, subjective terms.

According to this view, the subjective quality of the word, self, cannot be eradicated. Therefore, the word can never be objectively defined. When I speak of self I can speak with immediacy only of my own self. It is impossible to think of self without including its reflexive nature, the self experiencing itself. I can experience only my self directly, no other. But, sadly, the possibility of self-deception is always there. In the most basic terms, even conscious experiences of the nature of my body are highly variable; the distortions of my body image are known to me through fantasy and dream; the distorted images of my body can be further inferred from my analysis (including my self analysis). In more general terms, what I think is my reputation or value to others may be reasonably accurate or wildly out of keeping with the judgments of peers. More tellingly, my introspective judgments of what I am may be much less veridical -- at least in some ways -- than the analytic judgments of another person, especially if he knows me well as an artist might or as a psychoanalyst might.

But even these more objective judgments have great limitations. There is no way for another person to experience exactly what it is like to be me, no matter how much he is shown or told. He can come closest to knowing what I am "really" like if he is able to know a great deal about me and at the same time experience what he knows empathically -- that is, if he is able to "put himself in my shoes" and imagine with feeling what it is like to be me. He is assuming -- with feeling -- that my self has resemblances to his own self. But, although we know that empathic capacities have always been essential for any successful psychoanalyst, empathy remains an insecure foundation on which to build knowledge. It is easy, even for the most experienced person, to think he is being empathetic when he is only deluding himself in one way or another, or confusing his own self with the self of another.

I understand that there is more to my self than I can consciously know at any one time. Any individual who has worked productively in analysis has experienced the recovery from repression of aspects of his or her self that previously would have been stoutly denied. Inwardly, all of us know that there are things that can never be experienced directly but in a sense belong to our selves -- and must be, at least in some way, connected with them. We know that because the evidence from within makes the inferences inescapable. Even what we call id and superego "belong" to the ego. Schilder (1950) discussed the extensions of the body image in the form of possessions. Outwardly, I know that I -- my "self" -- can possess all sorts of things and even come to believe in the possession of other people. William James (1869) included personal possessions in his very definition of self. And I might even confuse myself with my possessions. Going further, I might even confuse myself with other people. But if all goes reasonably well, and I am able to mature to some extent, I may learn that although relations with other people make up part of myself, the people themselves do not. And I may learn that possessions may simultaneously belong to myself and to other people. I may learn to share without confusing who is who among the sharers. All that I have been saying is that my "self" consists of a series of more or less patterned and consistent fantasies, and these do not consist of what I think of as "me" alone, but "me" in relation to other parts of myself, or "me" in relation to other things, or "me" in relation to other people.

Now what my observer might call my "self" means what he thinks my self is, consciously and unconsciously, to me. Or, he may be simply defining the location of an action, in the same way the locations of actions are defined by ordinary pronouns. Actually, what the observer means about this part of my mind, this part he knows empathically by comparing it with his own self, and as objectively as he can by making a series of more or less reliable observations as a natural scientist might -- informed by his theories, including his theory of the ego as a system. This is what would be preferable for him to refer to, speaking on a theoretical level, as ego (or my "I"), providing these words implied both constructions derived closely from clinical events, and more abstract inferences drawn from them and collective experience of many analysts.

According to Freud's later theory,² the only access to those parts of mind considered to be id and superego is by way of ego. When an observer sits back to think about me as an individual, if he does so in English, he is apt to call what he perceives an ego, an ego which is capable of experiencing itself, and which shows evidence, at times of conflict, of id and superego. If he thinks in German or French or Spanish, he will call the products of his subjective and objective experiences and observations das Ich, le Moi, and e1 Yo, respectively. When an observer is thinking about a person as a unique individual, he is thinking of the ego's inner, experience-near relations with itself and other persons, and with its interactions with other aspects of the mind, the id, superego, and representations of the world. When an observer seeks to generalize about the egos of his patients, and combines and compares his accumulated empirical observations with those of others -- that is, when he moves from the individual to the group of human minds, he or she is using "ego" truly in its sense as an abstraction defined by its functions.

Clearly, Freud intended the words for the macrostructures of the structural theory to be derived from ordinary, everyday words which have the uncodified wisdom of the culture behind them. He could not have expressed himself more clearly than in the quotation (1926b) included above. Thus the ego would be, as Freud said in 1933, "our very own 'I'. (p. 58, my translation). What we call the id would be the impersonal "It" within us -- motivational forces which at unrestrained times can amount to the terrible. What we call the superego might be -- although of course it won't be -- the "Over-I," or "Beyond-Me," the part of us that urges the subjection of certain personal interests to those of some version of the collective. We can only know intrapsychic life in terms of what can be experienced, in terms of what is phenomenological. We can only build theories of the deep structure of the mind, the non-experiential, the noumenal world, by making valid inferences derived from the collective experiences of many analysts working over a long time in similar modes. In summary, for Freud Das Ich meant -- and from my point of view ego ought to mean: a clinical construction closely related to what is thought to be the "self" of the subject, comprehended by the observer in analogy and/or in complementarily with his own self, and joined with that expanded, more "objective" knowledge of himself which might be termed personal ego knowledge (for examples, knowledge of his own accustomed unconscious reactions and characterological propensities) usually acquired from his own personal analysis, along with the related, abstract, systemic meanings shared by other analysts operating according to the same fundamental scientific assumptions.

It is a virtue, not a vice, that both sets of meanings reside in the same word; the particular is thereby related to the general. As LaPlanche and Pontales (1973) said, the exploitation of this sort of ambiguity contained in one word amount to embracing real complexity in contrast to the spurious ease of simplicity and "neatness."

CRITICISMS OF THE UNITARY VIEW OF EGO

The criticisms of the unitary view just expressed fall into three categories: practice, the quest for objectivity, and the need for systematic economic concepts. Each area of criticism has its validity; at present, decisions can only be based upon what is most useful heuristically. Just because Freud did not always distinguish his meanings of das Ich as self or system does not mean we should follow him, the argument goes. Unless a writer is extremely careful, it is easy to confuse his intentions: does he mean to be speaking of that which is phenomenologically observable in an individual, or that which is inferred abstractly about minds in general? The practices developed on the basis of ego psychology have vastly improved the quality of analytic communication. True enough, one can never completely understand the self of another, but psychoanalysis has never pretended to be an exact science. It would be impractical to try to introduce new translations replacing ego; it would be clumsy and unworkable to expect authors to always write, "phenomenal ego," when they meant that, and "abstract ego," when they meant that. Better to leave existing semantic practices alone, the argument goes. Self is so commonly used in English that it retains great heuristic value even left undefined.

The second line of criticism focuses upon the avowedly subjective nature of the arguments in favor of a unitary view of das Ich. Psychoanalysis has scientific difficulties enough, it is said, without focusing upon its inescapable subjectivity; such a focus can introduce the same ambiguities and uncertainties that reliance upon introspection alone entails. What is needed, the argument goes, is emphasis upon objective, empirical observations made by psychoanalysts. It goes without saying that there are problems with the "purity" of these apperceptions, but better to focus on the positive nature of collective observations, with the built-in corrective tendencies implied, than to focus upon ephemeral and unanswerable questions having to do with what the "self" is, really.

Finally, the economic problem having to do with what "energizes" the "self" and what energizes the "ego" (as abstract system only) remains with us, even if Freud did not choose to deal with it. How can we develop psychoanalysis scientifically, this argument asks, without dealing with internal inconsistencies in its fundamental economic theories?

SELF-AS-STRUCTURE IN INTERACTION WITH EG0 AS STRUCTURE

A number of authors (e.g., Levin, 1963; Mitterauer and Pritz, 1978) have urged the construction of an abstract self as a fourth microstructure of the mind. As an equal of ego, the self would "interact" with it, just as it would with the id and the superego. For a time, it seemed that Kohut might come to the same conclusions, but he disavowed this particular suggestion (1979). Most critics have insisted that there are no clinical observations which can be brought forward to support such a reorganization of our structural abstractions, and that the introduction of a fourth microstructure would provide no heuristic advantages.

OTHER, SOMETIMES "SUPERORDINATE" VIEWS OF THE SELF

Actual or presumed defects in the Freudian explanations of certain clinical phenomena have always been occasions for theoretical alternatives. Sometimes, these interests have directly or indirectly stimulated alterations in "mainstream" thought. And sometimes, separate approaches have grown up about them. One that was popular for a time was Erikson's (1956, 1968, 1974) interest in "identity," a term which is usually, but not always, seen as synonymous with self. The confusions that developed about the relationship of identity to ego (Glover, 1966), have now been subsumed by almost identical confusions about the relationship of self to ego. Abend (1974) has contributed a useful survey of the various ways the term has been used -- and the clinical confusions which can result. (See Chapter 35. IDENTITY)

As the popularity of "identity" waned, other formulations, primarily of a cognitive stamp, waxed. Among them are the considerations of the "representational world," begun by Sandler and Rosenblatt (1962), the attempts to develop various cognitive explanations to augment or replace standard metapsychology, the search for alternative explanations of the mental apparatus which utilize systems theory or computer science as models, influences from those who see psychoanalysis as an "hermeneutic" activity opposed to empirical scientific work, rather than in consonance with it.

The popularity of two competing (but constantly changing) views of the "self" took center stage in the late 1970's. Kernberg (1976, 1982), at first heavily influenced by Kleinian thought, addressed his theoretical speculations to problems of the more disturbed patients, "borderline" and "narcissistic" personalities. Kohut (1971, 1977, 1978, 1982) was also interested in "narcissistic" personalities, but defined them differently, on the basis of characteristic transference (or transference-like) reactions. The specific assumptions of the rival approaches will not be addressed here. See, for examples, Calef & Weinshel (1979) and Rangell, (1982) for critiques of Kernberg's system; see, for examples, Curtis (1983) and Treurniet (1983) for critiques of Kohut's and his followers' work. Each eventuated in a system which postulated a "superordinate self" (despite Kohut's expressed disavowal, mentioned above). Each also postulated an entity called the "grandiose self," but the two versions were defined so differently they were not relatable.

According to Kernberg (1974), the grandiose self represents a pathological condensation of three other inferred structures: the "real self," the "ideal self," and the "ideal object." Kohut (1977) says that the grandiose self is one pole of the "self (deliberately left undefined). Kernberg (1982), urges two incompatible points about the relations between self and ego. In one, he decries the translated division of Freud's das Ich into ego and self. In the other, he postulates a mature, overarching "superordinate self." His reasoning is not explained, nor is it self-evident. He states that the self "is an ego structure that originates from selfrepresentations ... It is, in short, an ego function and structure that evolves gradually from the integration of its component self-representations into a supraordinate structure that incorporates other ego functions ...and leads to the dual characteristics implied in Freud's Ich" (P. 905).

Kohut's early work having to do with the self (1966, 1968, 1971) centered on a modified economic theory (libido was seen in terms of two developing streams of motivational forces, one related to objects, the other related to the individual's self love and narcissistic idealizations of others). The later work (1977) postulates a separate line of development of the "bipolar self" which is completely independent from drives and their attendant intrapsychic conflicts. The resulting "self psychology," a version of "object relations psychology," is discussed in Chapter 33, SELF PSYCHOLOGY, and will not be detailed here.

Essentially, the resulting theory postulates a development (and disorders of development) of a self which first relates to objects (as seen by an outside observer) and which might more accurately be termed "selfobjects," inasmuch as they exist in terms of the needs of the self from the inner viewpoint of the individual. If needs to be admired and loved ("mirrored") by the selfobject, usually (to the external observer) the parent, or if the needs of the self to idealize selfobjects are not met, specific pathologies develop. These are to be distinguished from pathologies related to drives and conflicts. The resulting techniques of treatment depend on a vastly expanded concept of empathy.

It is to be noted that, although Ornstein (1981) implies that Kohut had a clear-cut conception of "self" from the beginning, Kohut nevertheless publicly refrained from explicit definitions (Wolf, 1985). Kohut believed that a premature definition would inhibit and restrict further development of the self-psychological concepts. As followers and non-followers of Kohut's thought acquire further experience, no doubt the clinical validity of the concepts will be determined, and almost certainly, more explicit definitions of this meaning of self will be forthcoming. On the other hand, the evidence for or against Kernberg's theories, including his superordinate self, can hardly come from clinical evidence inasmuch as they are made up of interlocking deductions. At this point most psychoanalysts have withheld judgment, or been openly dubious about the claims of either system (Curtis, 1983; Blum, 1982; Rangell, 1982; and Richards, 1982).

The same can be said for two other theories which postulate the centrality of self, both on the basis of dissatisfactions with existing metapsychological formulations. George S. Klein (1976), in an effort to rid psychoanalysis of metapsychology altogether, proposed a "self-schema," which would be able to have both "human" and "systemic" attributes. However, as Richards (1982) points out, once the self-schema is defined as a psychic apparatus, something like a metapsychology becomes re-inserted. John Gedo (1979) makes the epigenesis of a "self-organization" central. He focuses on a hierarchy of values and personal aims. The importance of biological motivations become integrated with goals and values. Technical recommendations based on this reorganization of theory await the tests of time.

It may be that the tendency to reify and elevate self and identity to "superordinate" or "supraordinate" levels of abstraction is primarily a phenomenon of English-speaking analysts. But, as Glover (1966) suggested, the tendency began at least with Jung. Some colleagues wish to locate some sort of "primal self" at the beginning, which might open up like a paper flower in a glass of water; or to discern over-riding "identity themes," beginning in the earliest mother-infant relationship in which, it is thought, the mother's unconscious might serve as a sort of cast (Lichtenstein, 1963, 1977). There are universal wishes to discover "core" meanings of human life, both in its beginnings and its ultimate purposes. There is nothing to criticize in such wishes, or theories derived from them, unless they invade the core propositions of psychoanalysis as a discipline.

CONCLUSIONS

The literature on the self is massive and confusing. Terms are not always concepts; sometimes they merely cover vacuums. A redundancy exists: self, identity, identity themes (along with mysterious hybrids: ego identity and self identity), variously mean the individual, the mind (phenomenally or noumenally), or even something like a metaphysical fate, as in identity themes -- enough to fill many volumes. Too often, attempts to clarify the Freudian system, or essentially replace some part of it, extirpate its core, its basic concepts.

This chapter takes up the beliefs of the majority of mainstream analysts, in America at any rate, that ego should be defined only as an abstraction which includes a cohering set of psychic functions, and that "self" is a different entity. An opposing argument among mainstream analysts (including myself) has it that the ego should be understood as a term exactly comparable to Freud's term, das Ich, a concept with a range of meanings, from constructs based on clinical observations to high levels of abstraction. Further propositions have been advanced by more dissident analysts: that "self" should be seen as the fourth microstructure of the structural theory; or that "self" should be seen as "superordinate" to the id-ego-superego system. Some of the criticisms of each point of view were advanced.

 ENDNOTES

1. Before the introduction of the concept of narcissism, Freud postulated a duality of psychic energies, libido (the sexual, species-preservative instincts) and ego-instincts (the self-preservative instincts), which were manifested, Freud thought, by "egoism." With the introduction of theories of narcissism, the ego also was presumed to be cathected by libido. For a period, the self-preservative instincts and egoism rarely appeared in the literature; it was not until the 1920's that Freud's last duality of instincts appeared in the form of Eros and the Death Instinct. Nevertheless, Freud continued to occasionally refer to the older terms. 

2. Earlier, Freud spoke of unconscious to unconscious communications. 

REFERENCES

ABEND, S. (1974). Problems of identity: theoretical and clinical applications. Psychoanal. Q., 43:606-637. 

ARLOW, J. and BRENNER, C. (1964). Psychoanalytic Concepts and the Structural Theory. New York: Int. Univ. Press. 

BERES, J. (1981). Self, identity, and narcissism. Psychoanal. Q., 50:515-534. 

BETTELHEIM, B. (1983). Freud and Man's Soul. New York: Knopf. 

BLUM, H. (1982). Theories of self and psychoanalytic concepts: discussion. J. Amer. Psychoanal. Assn., 30:959-978. 

BOESKY, D. (1983). The problem of mental representation in self and object theory. Psychoanal. Q. 52:564-583. 

BRENNER, C. (1980). Metapsychology and psychoanalytic theory. Psychoanal. Q. 49:189-214. 

CALEF, V. & WEINSHEL, E. (1979). The new psychoanalysis and psychoanalytic revisionism. Book review essay on Borderline Conditions and Pathological Narcissism. Psychoanal. Q., 48:470-491. 

CURTIS, H. (1983). Review of The Search for the Self: Selected Writings of Heinz Kohut 1958-1970. Edited by Paul H. Ornstein. New York: Int. Univ. Press. J. Amer. Psychoanal. Assn., 31:272-285. 

ELLENBERGER, H. (1970). The Discovery of the Unconscious. The History and Evolution of Dynamic Psychiatry. New York: Basic Books. 

ERIKSON, E. (1956). The Problem of ego identity. J. Amer. Psychoanal. Assn., 4:56-121. 

ERIKSON, E. (1968). Identity. London: Faber and Faber. 

ERIKSON, E. (1974). Dimensions of a New Identity: The 1973 Jefferson Lectures in the Humanities. New York: W.W.Norton. 

FREUD, A. (1936). The Ego and the Mechanisms of Defence. New York: Int. Univ. Press. 

FREUD, S. (1895). Project for a scientific psychology. S.E., 1:295-397. 

--------------- (1914). On narcissism: an introduction. S.E., 14:73-102. 

--------------- (1915). Instincts and their vicissitudes. S.E., 14:117-140. 

--------------- (1923). The ego and the id. S.E., 19:12-66. 

--------------- (1926a). Inhibitions, symptoms and anxiety. S.E., 20:87-172. 

--------------- (1926b). The question of lay analysis. S.E., 20:183-250. 

--------------- (1933). New introductory lectures on psychoanalysis. S.E., 22:5-182. 

FRIEDMAN, L. (1980). The barren prospect of a representational world. Psychoanal. Quart., 49:215-233. 

GEDO, J. (1979). Beyond Interpretation. New York: Int. Univ. Press. 

GILL, M. (1963). Topography and Systems in Psychoanalytic Theory. New York: Int. Univ. Press. 

GLOVER, E. (1966). Metapsychology or metaphysics: a psychoanalytic essay. Psychoanal. Q., 35:173-190. 

GROSSMAN, W. (1982). The self as fantasy: fantasy as theory. J. Amer. Psychoanal. Assn., 30:919-938. 

HARTMANN, H. (1939). Ego Psychology and the Problem of Adaptation. New York: Int. Univ. Press. 

-------------------------(1950). Comments on the psychoanalytic theory of

the ego. Psychoanal. Study Child., 5:74-96. 

-------------------------(1953). Contribution to the metapsychology of schizophrenia. Psychoanal. Study Child., 8:177-198. 

------------------------ (1955). Notes on the theory of sublimation. Psychoanal. Study Child., 10:9-29. 

------------------------ (1956). The development of the ego concept in Freud's work. Int. J. Psychoanal., 37:425-438. 

JACOBSON, E. (1964). The Self and the Object World. New York: Int. Univ. Press. 

---------------------- (1983). The structural theory and the representational world: developmental and biological considerations. Psychoanal. Quart., 52:543--563. 

JOFFE, W. and SANDIER, J. (1967). Some conceptual problems involved in the consideration of disorders of narcissism. J. Child Psychother., 2:56-66. 

KERNBERG, O. (1974). Barriers to falling and remaining in love. J. Amer. Psychoanal. Assn., 22:486-511. 

------------------------(1976) Object Relations Theory and Clinical Psychoanalysis. New York: Aronson.

----------------------- (1982). Self, ego, affects and drives. J. Amer. Psychoanal. Assn., 30:893-918. 

KLEIN, G. (1976). Psychoanalytic Theory. New York: Int. Univ. Press. 

KOHUT, H. (1970). Moderator's opening and closing remarks to discussion of D. C. Levin's "The self: a contribution to its place in theory and technique." Int. J. Psychoanal., 51:176-181. 

---------------- (1971). The Analysis of the Self. A Systematic Approach to the Psychoanalytic Treatment of Narcissistic Personality Disorders. New York: Int. Univ. Press. 

---------------- (1977). The Restoration of the Self. New York: Int. Univ. Press. 

----------------- (1978). The Search for the Self. Selected Writings of Heinz Kohut.

Ed. P. Ornstein. New York: Int. Univ. Press. 

LAPLANCHE, J. and PONTALES, J.-B. (1973). The Language of Psychoanalysis. New York: W.W. Norton. 

LEVIN, D. (1969). The self: a contribution to its place in theory and technique. Int. J. Psychoanal., 50:41-51. 

LICHTENSTEIN, H. (1963). The dilemma of human identity: notes on self-transformation, self-observation and metamorphosis. J. Amer. Psychoanal. Assn., 11:173-223. 

LICHTENSTEIN, H. (1965). Towards a metapsychological definition of the concept of self. Int. J. Psychoanal., 46:117-128. 

LOEWALD, H. (1971). On motivation and instinct theory. Psychoanal. Study Child,. 26:91-128. 

--------------------- (1973). Review of The Analysis of the Self. A Systematic Approach to the Psychoanalytic Treatment of Narcissistic Personality Disorders, by H. Kohut. Psychoanal. Q., 42:441-451. 

LOEWENSTEIN, R. (1940). The vital or somatic instincts. Int. J. Psychoanal., 21:377-400. 

MAHLER, M. & McDEVITT, J. (1982). Thoughts on the emergence of the sense of self, with particular emphasis on the body self. J. Amer. Psychoanal. Assn., 30:827-848. 

MAHLER, M., PINE, F. & BERGMAN, A. (1975). The Psychological Birth of the Human Infant. Symbiosis and Individuation. New York: Basic Books. 

MITTERAUER, B. and PRITZ, W. (1978). The concept of the self: a theory of self-observation. Int. Rev. Psychoanal., 5:179-188. 

ORNSTEIN, P. (1981). The bipolar self in the psychoanalytic treatment process -- theoretical considerations. J. Amer. Psychoanal. Assn., 29:353-376. 

ORNSTON, D. (1982). Strachey's influence: a preliminary report. Int. J. Psychoanal., 63:409-426. 

RANGELL, L. (1982). The self in psychoanalytic theory. J. Amer. Psychoanal. Assn., 30:863-892. 

RICHARDS, A. (1982). The superordinate self in psychoanalytic theory and in the self psychologies. J. Amer. Psychoanal. Assn., 30:939-958. 

SANDLER, J. and JOFFE, W. (1969). Towards a basic psychoanalytic model. Int. J. Psychoanal., 50:79-90. 

SANDLER, J. and ROSENBLATT, B. (1962). The concept of the representational world. Psychoanal. Study Child, 17:128-145. 

SCHAFER, R. (1968). Aspects of Internalization. New York: Int. Univ. Press. 

SCHAFER, R. (1973). Concepts of self and identity and the experience of separation-individuation in adolescence. Psychoanal. Q., 42:42-59. 

SCHILDER, P. (1950). The Image and Appearance of the Human Body. New York: Int. Univ. Press. 

SPRUIELL, V. (1975). Three strands of narcissism. Psychoanal. Q., 44:577-595. 

-------------------- (1979). Object relations theory: clinical perspectives. J. Amer. Psychoanal. Assn., 27:387-398. 

---------------------(1981). The self and the ego. Psychoanal. Quart., 50:319-344. 

STRACHEY, (1961). Editor's Introduction. S.E., 19:7-8. 

TREURNIET, N. (1983). Psychoanalysis and self psychology: a metapsychological essay with a clinical illustration. J. Amer. Psychoanal. Assn. 31:59-100. 

WOLF, E. (1985). Personal Communication.