The Ideal of the Anonymous Analyst and the Problem of Self-Disclosure

What constitutes understanding in the clinical situation? What constitutes authority? How are understanding and authority managed in the relationship between analyst and analysand? These are questions very much under discussion right now in psychoanalytic circles. A number of fascinating philosophical and epistemological issues are implicated; but they devolve, for the practitioner, onto everyday choices concerning technique, perhaps none more crucial than the problem we usually take up under the heading of the analyst's self-disclosure.

I think we've come to the point at which we need to review the way we conceptualize self-disclosure by an analyst, and our assumptions about the effect of such self-disclosure upon the progress of the psychoanalytic investigation. I will propose, in remarks to follow, that our prevailing conceptions about analytic anonymity serve different and less constructive purposes than we have thought, purposes that bear directly on the issues of how understanding is arrived at in analysis, and how authority in the treatment relationship functions. I will suggest that we can benefit from a more systematic consideration of useful forms of self-disclosure by the analyst that currently have to be bootlegged in and around the edges of a theory of technique which in principle discourages them; and I will outline a logic of self-disclosure that I think may be helpful. My purpose, I should say, is neither to substitute a new set of categorical prescriptions for the old ones, nor to advocate for spontaneous self-revelation by the analyst, but rather to enlarge our clinical repertoire by systematically reviewing and revising certain of our technical guidelines. I intend to argue for a greater degree of freedom, and ultimately for more self-consciousness, in making our choices concerning what to say about ourselves to our patients.

I believe it is generally accepted that analytic work is facilitated when the analyst is able to maintain a posture of maximum anonymity. Of course, contemporary analysts tend to be "flexible" in their application of the principle of anonymity. A human demeanor for the analyst is widely recommended nowadays and rigid hyper-formality frowned upon. Many feel that, as a practical matter, judicious self-revelation by the analyst under certain circumstances can be "the least evil" (e.g., Hoffman, 1994). For example, I think there is probably consensus that when an analyst has "made a mistake" -- let's say gotten angry and insulted a patient -- the analyst should admit it.

Effective clinical analysts tend not to follow the principle of anonymity with absolute strictness; they frequently set it aside, or interpret it idiosyncratically, somehow or other find a way to work around it in order to get the job done with their patients. Self-disclosure by the analyst may be seen as necessary because of particular circumstances (the patient is a child, an adolescent, is especially disturbed) or conceptualized as part of something other than analytic work per se (establishing the therapeutic alliance, maintaining engagement or rapport); but however they arise, departures from the principal of analytic anonymity don't alter its central place in our prevailing theories of technique. On the contrary, they preserve the principle of analytic anonymity by sparing it full accountability.

The premise that self-disclosure by an analyst burdens the analytic work remains intact and influential: we are directed by theory to subtract our personalities from the analytic situation, as far as possible, in order to leave our patients the blankest screen available upon which to project their fantasies. We may have our doubts about the ideal of the-analyst-as-reflecting-mirror and want to feel we have left it behind, but we have not yet really replaced it; we've just made it more user-friendly by not taking it entirely seriously.

A major difficulty with the technical injunction against self- disclosure is that anonymity for the analyst is impossible -- not only complete anonymity, but any anonymity at all. This is a radical statement, I know, but I don't think it is an exaggeration.

Every intervention hides some things about the analyst and reveals others (see, e.g., Chused, 1990; Greenberg, 1991); and every decision not to intervene communicates something, since patients tend to be quite aware of analysts' silences. I think we commonly make the mistake of thinking that when we impose inhibitions upon ourselves in the clinical situation (for example, keep quiet instead of yielding to the temptation to make a potentially seductive remark to a patient) that we reduce the degree to which the analyst's personality makes itself felt. Actually, we have only altered the manner in which the analyst's personality makes itself felt. Careful examination shows that any way an analyst decides to deal with his or her emotional responses is consequential. Since that is so, the question becomes not whether to disclose, but how to manage the unavoidable condition of constant disclosure. In my view, to suggest that an analyst can minimize communication of his or her idiosyncratic psychology -- emotional reactions, personal values, constructions of reality, and the like -- is to advocate pursuit of an illusion.

Elsewhere (Renik, 1993) I've explained in detail why it seems to me that expression in action of an analyst's affectively charged involvement always precedes his or her awareness of it, making analytic technique irreducibly subjective. I think we can recognize only after the fact how our personal reactions have been manifesting themselves. For present purposes, I would emphasize that an analyst's personality is constantly revealed, in one form or another, through his or her analytic activities. We can put our hands over our eyes, if we want; but we won't disappear.

Very much to the point is an observation Singer (1977) makes:

"...analysts often appear peculiarly reluctant to comment insightfully and incisively on their clients' communications... they seem fearful that their insight would make self-evident that the analyst, too, 'has been there,' at least at some point in his life. Their empathic grasp, they correctly sense, could betray pointedly that the basic precondition for empathic communion is given, that is, personal knowledge of the experience under scrutiny... what analysts so fondly think of as interpretations are neither exclusively nor even primarily comments about their clients' deeper motivations, but first and foremost self-revealing remarks." (p. 183, my italics).

My impression is that Singer is right. What he says underlines not only that anonymity for the analyst is a fiction, but that the need to pretend to anonymity can have a constraining and deforming effect upon an analyst's clinical efforts. What it is best for an analyst to say or not say about him- or herself to a patient at any given moment remains an important and consequential decision. However, recognition that all of an analyst's analytic activity involves one form or another of self-disclosure obliges us to reconsider what we mean by self-disclosure and to think systematically about what types of information about the analyst are useful to communicate to a patient.

I want to mention in this connection an important influence that has had a liberating effect upon analytic technique, but has, ironically, at the same time perpetuated the ideal of the anonymous analyst: namely, the conception of the analytic process advanced by a number of theorists who make use in a particular way of the concept of projective identification. Bollas (1987) articulates the conception very clearly when he writes:

"...for differing reasons and in various ways, analysands re-create their infantile life in the transference in such a determined and unconsciously accomplished way that the analyst is compelled to re-live elements of this infantile history through his countertransference, his internal response to the analysand." (p. 200).

According to this point of view, the analyst's subjectivity in effect presents no technical problem because it is inconsequential: such is the patient's power to determine the analyst's experience that the analyst's individual psychology is overridden; the analyst is, therefore, for all practical purposes, rendered anonymous.

"To find the patient," Bollas says, "we must look for him within ourselves." (p. 202). The analyst is presented with appealing modesty as a vessel for transference, a "potential space" within which the patient can "live infantile life anew." (p. 200). However, it is also true that the analyst is assumed to approach becoming a perfect observing instrument, transcending his or her idiosyncracy through submission to a powerful analytic process. The analytic relationship is envisioned very much in the tradition of Freud's (1915) romantic pronouncement, "It is a very remarkable thing that the Uncs. of one human being can react upon that of another, without passing through the Cs." (p. 184).

Yet we cannot avoid asking: when the analyst looks inside, how is it that the analyst sees a re-creation of the patient's infantile life, rather than the analyst's own experiences, independently determined to a significant extent by his or her own individual psychology? Even if we grant that a patient may strive to elicit in an analyst a duplication of the patient's life struggles, why does the analyst's subjectivity not constitute a powerful obstacle to faithful re-creation? Bollas explains that an analyst must be well-analyzed enough to allow himself or herself to regressively experience and contain the countertransference. If he or she can do that, the analyst's "neutrality" creates a "frame," a "dream screen" against which the transference is played out (p. 201).

Here is the familiar core conception once again -- old wine in a new bottle: the ideal of the reflecting mirror remains intact, but has been relocated; now the analyst, rather than the patient, gazes into it. The notion of a patient "out there" who can be studied and known by an anonymous, objective analyst is exchanged for the notion of a patient "within" who can be similarly studied and known.

What is of special interest about this particular conception is that while it retains an idealized picture of the analyst as anonymous observer, it also encourages analysts who subscribe to it to be rather more freely expressive in their interventions than others. Thus, Bollas writes:

"...it is crucial that the clinician should find a way to make his subjective states of mind available for the patient...even when he does not yet know what these states mean... analyses rarely proceed with such clarity that the clinician knows in statunascendi what and whom he is meant to become..." (pp. 200-201).

If an analyst believes that his or her personal reactions are being controlled -- in ways and for reasons that may not yet be entirely clear -- by a patient's manner of participation in the analytic situation, the analyst need not be reluctant to report those responses to the patient. The analyst will not feel that his or her anonymity is compromised in so doing, because the reports will not be understood, ultimately, to constitute personal self-disclosures -- they are merely descriptions by the analyst of the patient within himself or herself in raw form, reflections of not fully digested transference material.

For my own part, I do not entirely agree with this understanding of the transactions that take place between analyst and analysand. We have to be careful not to fall into teleological assumptions as we try to make sense of clinical events. I think that some patients, sometimes, do try to get their analysts to feel what they feel, or have felt; and in other instances, an analyst does come to have experiences very similar to ones his or her patient is warding off -- though not because of purposeful instigation by the patient. I find the concept of projective identification to have great value, but I think it is sometimes used in a mechanistic, even magical way that describes the movement from one person to another of thoughts and feelings conceived of as concrete objects. When this is done, it seems to me, a fantasy that is sometimes entertained by one, or even both, members of an analytic couple, each in his or her own way, becomes confused with an accurate description of events.

One consequence of the confusion is that the subjectivity of the analyst's perceptions is disavowed, and an undue authority for the analyst as observer preserved. Nonetheless, I believe that the influence upon technique of such a view has a salutary aspect, inasmuch as the analyst is encouraged to state his or her experience frankly and explicitly to the patient. I regard this as a good thing happening, in part, for the wrong reasons!

I hasten to add that many individual analysts manage, through aspects of their personal styles, to descend from the elevated position in which their theories of technique would place them. When Bollas, for example, describes his clinical work, he conveys an awareness of the fallibility of his formulations and a respect for the epistemological privacy of his patients that is certainly communicated to them. Still, we are best off with a theory of technique that does not have to rely on a given analyst's personal modesty to undo its unfortunate implications. The problem of the analyst's position as an objective authority, it seems to me, is hardly confined to certain variants of Kleinian thought.

Among American analysts, the usual line of reasoning that argues against deliberate self-disclosure by the analyst is based on a distinction made between reality and fantasy. The mind is conceptualized as in Arlow's (1969) well-known analogy, something like a screen upon which images are being projected from both within and without. If the contributions from without can be reduced, if very little about the analyst is revealed, then there is less interference with the patient's identification of his or her internally-generated imagery; whereas, to the extent that the patient is given information about the analyst, it permits the patient to experience his or her perceptions of the analyst as if they were simply appraisals of incoming sensory data, and the patient is not as ready to acknowledge the influence of wishful thinking and unconscious preconceptions. In other words, as it is often put: the more a patient is presented with realities about the analyst, the harder it is for the patient to acknowledge his or her transference fantasies.

Yet we know that every analytic encounter presents the patient with myriad "realities" about the analyst. Furthermore, the things an analyst "really" does when the analyst strives for anonymity are just as likely to correspond with a patient's crucial unconscious expectations as are the "realities" presented by purposeful self-disclosure on an analyst's part. If an analyst doesn't answer questions, remains silent much of the time, and never reports personal feelings or opinions, we are familiar with how easily this can be experienced, by some patients, as confirmation of a belief that the analyst is sadistic and withholding, or needing to be in control for competitive reasons; on the other hand, the very same reserve and suppression of self-expression on the analyst's part is, by other patients, just as easily construed to reflect the analyst's selfless devotion and used to give credence to profound magical hopes and wishes.

Whatever an analyst does, he or she is constantly dumping grist into the proverbial mill; and the notion that the quantity of grist will be limited if the analyst pursues a policy of behavioral minimalism and maintains an impersonal demeanor is received wisdom that seems to me to be contradicted by what our collective clinical observations actually indicate. My own experience suggests to me that whether I choose to comment on my patient's apparent submissiveness to his wife, or tell him that I find his description of his relationship with his son very touching, or attend a dinner party at which he is present, my patient will arrive at certain conclusions about me by a process in which observation and inference are inextricable. What gives us reason to say categorically that a greater quantity of information, or a certain kind of information, provides more "reality" and less opportunity for the generation of "fantasy," than another?

The articles that have been written about extra-analytic encounters and other unusual interactions (e.g., those required by illness in the analyst) certainly indicate that when ordinarily avoided forms of self-disclosure are thrust upon an analyst, subsequent analytic investigation of them can be extremely productive (see, e.g., Abend, 1986; Ganzarain, 1991). Why do we assume that these are special circumstances, that they burden an analysis, and that the yield we are able to take from exploration of them when they occur essentially represents damage control? If we look at the results actually reported in our literature, we see that all sorts of analyst-analysand interactions, from the most conventionally "interpretive" to the most obviously revealing about the analyst, are occasions for productive analysis (and this does not include those accounts of analytically beneficial, unorthodox encounters that are talked about informally, but never get written up because they don't square with existing theory). I think we have to admit that a blanket principle of analytic anonymity does not, in fact, help us determine which forms of self-disclosure by an analyst are likely to oppose and which facilitate analytic investigation. In my view, the distinction between reality and fantasy best refers to a judgment each individual makes concerning his or her various experiences as they occur. To regard certain of an analyst's behaviors as inherently "more real" than others, and therefore more foreclosing of "fantasy," is to reify a set of phenomenological concepts.

It is interesting that the ideal of anonymity for the analyst has not received more explicit challenge within psychoanalytic circles, inasmuch as there are theories of analytic process that would seem to argue implicitly against traditional ideas about self-disclosure and technique. If an analyst places primary emphasis on the importance of healing interactions within the treatment relationship, as opposed to the pursuit of insight, there is no reason for the analyst to strive for a posture of anonymity. Analytic anonymity is not intended to generate new experiences with a new object; it is a strategy designed to maximize conscious scrutiny of a patient's previously unconscious mental life.

For a self psychologist, trying to effect empathic repair of deficits caused by narcissistic injury, self-disclosure per se should not necessarily be contraindicated; likewise, for a control mastery analyst, who is concerned with passing a patient's tests so as to disconfirm the patient's pathogenic beliefs. Ehrenberg (1984), for example, says: "The analyst has to be instrumental in generating a new experience of a different order. There are certain ways of engaging the patient that will encourage and facilitate..." (p. 23). According to the conceptions of technique that follow from analytic theories that see the treatment relationship as curative, an analyst at various moments in the treatment wants to be revealed to the patient as having one attitude or another. Indeed, my impression is that analysts who subscribe to such theories are by and large not quite so meticulously concerned with trying to remain anonymous to their patients. Nonetheless, it seems to me, analysts overall tend to avoid deliberate self-disclosure, even in the absence of specific theoretical justification for the avoidance. I would say that among analysts of all theoretical orientations, there is significant reluctance to completely abandon a posture of analytic anonymity.

Why the pervasive tendency to avoid self-disclosure? If we want to try to understand why anonymity for the analyst has endured as a technical ideal, it may be instructive to consider the consequences of maintaining the ideal. Of course, what is achieved thereby is not actual anonymity for the analyst, but a pretense of anonymity. The analyst has the illusion that he or she can remain relatively anonymous in the analytic situation, and via communication of conviction about this illusion, invites the patient to subscribe to it. The result is a collusion in which both analyst and patient disavow revelations of the analyst's subjectivity in the treatment situation and the patient's capacity to perceive them. A kind of folie- -deux is set up at the heart of the treatment relationship, encouraged by our theory of technique. Hoffman (1983), in discussing this situation, speaks aptly of "the myth of the naive patient." I want to focus on the obverse side of the myth -- the image of the analyst that is promoted by a pretense of anonymity.

The pretense of anonymity is a cloak worn by the analyst when pictured as an authoritatively objective observer, able to transcend his or her subjectivity in the treatment situation. Conviction on an analyst's part of being able to achieve authoritative objectivity, even to a relative degree, constitutes a very powerful self-idealization; and it is this idealization of the analyst in which the patient is encouraged to participate. Cooper (1993) describes an aspect of the idealization when he speaks of reluctance to acknowledge the analyst's "interpretative fallibility." Denial of the analyst's interpretative fallibility can be discerned not only in some Kleinian conceptions of technique, as I mentioned earlier, but in mainstream "ego psychology" ones as well. For example, consider what is often called "the use of external reality as a defense." Discussing how this defense should be addressed technically, Inderbitzen and Levy (1994) write:

"Reality intrusions...interfere with the analysand's capacity for self-observation, especially of unconscious id and superego pressures. It is our clinical impression that patients turn their attention to...realities... in order to defend themselves against observing and fully experiencing intrapsychic pressures. It is here that the analyst intervenes..." (p. 777).

One wonders how, according to these authors, an analyst is supposed to know when a patient is attending to "reality" rather than to "unconscious phenomena." The question, "reality according to whom?" apparently does not arise. How does an analyst know when a patient is fully experiencing "id and superego pressures"? Clearly, Inderbitzen and Levy are not troubled by claiming authority for an analyst's judgments over a patient's about the patient's own experience. Inderbitzen and Levy assume that there is only one reality, objectively determinable by the analyst. Since they see the analyst as arbiter of reality, they also see the analyst as having the responsibility of minimizing intrusions of reality into the treatment. They believe that by refraining from self-disclosure the analyst can avoid intruding his or her "real" self into the analytic situation.

An important corollary of the principle of analytic anonymity is the widely-subscribed-to technical premise that it's not advisable to "reality test" with patients: i.e., it is thought that for an analyst to explicitly state his or her own view of reality constitutes a personal disclosure on the analyst's part that tends to foreclose a patient's exploration of his or her own view. I believe that when put into action, this premise invites idealization of the analyst. By feeling the need to withhold his or her views of reality so as not to influence a patient, an analyst conveys the conviction that his or her views are, in fact, potently authoritative. The analyst communicates the expectation that if a patient were to be exposed to the analyst's views, the patient would no longer be willing or able to think for himself or herself. Thus, any tendency on the patient's part to award the analyst undue authority as an arbiter of reality, instead of being considered unnecessary and held up for scrutiny, is implicitly endorsed as unavoidable.

By contrast, an analyst who regards his or her own constructions of reality as no more than personal views to be offered for a patient's consideration has no reason to avoid stating them explicitly. Reality testing, if we want to call it that, takes the form of interventions in which the analyst presents a point of view different from the patient's by saying, essentially, "Here's what I see. Here's what it suggests to me. What do you see and what does it suggest to you?". In this vein, Bollas (1987) describes how Winnicott, when making interpretations, treated his own ideas about reality as "subjective objects placed between analyst and patient."

I have been noting that the principle of analytic anonymity encourages idealization of the analyst as an authoritative observer of reality within the treatment situation. Inasmuch as this is the outcome, we must consider that it is the desired outcome. It may be painful for us to acknowledge that a longstanding, fundamental principle of analytic technique is actually designed to promote irrational overestimation of the analyst, but we cannot really be surprised. After all, Freud was unapologetic about cultivating idealization of the analyst in the service of the treatment. His idea that "unobjectionable positive transference" should be used to facilitate the "overcoming of resistance" is well known. Perhaps we never really abandoned it. The desire to maintain a distinct identity for psychoanalysis certainly led analysts who succeeded Freud eventually to reject the idea that suggestion based on the authority of the doctor is a crucial part of our clinical method. To think otherwise would have been to admit that the mechanism of action of clinical psychoanalysis has much in common with all sorts of other psychotherapies -- not to mention with hypnosis, shamanistic healing rituals, and the like.

In order to be sure that psychoanalysis is not just another therapy based on covert omnipotent fantasies about the therapist, we have developed an increasingly sophisticated theory of the analysis of transference over the years. The trend has been toward a more and more radical examination, a de-construction of the analyst as the sujet suppos‚ savoir (the one who is supposed to know), to borrow Lacan's felicitous phrase. I would say, however, that at the same time, maintenance of the ideal of the anonymous analyst has provided a powerful, unacknowledged counter current. It may be that in our eagerness to believe we have been successful in devising a method for analysis of transference, including "unobjectionable positive transference," we have disavowed the central way in which we have perpetuated and leaned upon the very phenomenon we thought to be eliminating: a policy of "non-disclosure" and maintenance of the ideal of an "anonymous" analyst has permitted us implicitly to solicit and accept idealization even while we are ostensibly involved in ruthless analysis of it. In order to adopt a technical stance that truly seeks to de-construct his or her undeserved authority, an analyst has to be confident that he or she can operate without it, can offer a cure that is not, ultimately, based on suggestion. How common is that confidence, really, among analysts?

I think we can all agree that idealization of the analyst by the patient is a crucial, useful phase in certain analyses, perhaps to some degree in all analyses. Idealization of the analyst is not, in and of itself, something to be avoided or suppressed; it is a phenomenon to be understood when it arises. In fact, if an analyst cannot tolerate being idealized, this can interfere with the necessary unfolding of the treatment relationship and prohibit important analytic work. Also, it is true that the psychoanalytic situation permits analysts to treat their patients better in some ways than analysts treat other people in their lives -- what might be called the "actual" idealization of the analyst (see Hoffman [1994] on idealization in interactional terms) leading to earned authority. However, while idealization of the analyst initiated by a patient out of his or her needs, or actual ideal behavior by an analyst arising from the structure of the analytic relationship, is not necessarily counterproductive, when an analyst solicits, consciously or unconsciously, idealization and unearned authority, it has significantly problematic consequences.

An analyst's wish to be therapeutically effective via disavowed authoritative suggestion can dovetail with other wishes. We know that being idealized as an authority can be personally gratifying for an analyst, can afford the analyst protection from anxieties that he or she otherwise experiences in interpersonal interchanges. Evidence of this is the unhappy fact that we not infrequently see analysts maintaining an "analytic" stance outside the clinical setting: with patients after analysis has been concluded (justified as necessary in case the patient might want to return to treatment), in supervision, and even in ordinary social situations. At the same time, anonymity is often relinquished more easily with those ex-analysands who become analysts themselves. Perhaps this is because shared values, not to mention identification and the willingness to enter into a hierarchy at a subordinate position, promise that idealization of the analyst will be maintained.

What is the remedy? Certainly the whole trend of the past ten years or so toward a theory of technique based on an intersubjective conception of the analytic situation has begun to treat analytic anonymity as a myth and to address the idealizations promoted by the myth. For example, Hoffman (1983) emphasizes the importance of recognizing that an analyst's personality is always expressed behaviorally in the here and now. Hoffman defines transference manifesting in the treatment situation in terms of a patient's need, for unconscious reasons, to selectively attend to only one plausible interpretation among many possible plausible interpretations of an analyst's conduct. Thus, the assumption that an analyst can be anonymous and can function as privileged interpreter of a patient's experience ("realistic" versus "distorted by transference") is rejected. Instead, the patient is recognized to be as much a legitimate interpreter of the analyst's experience as vice versa. Aron (1991) illustrates the clinical implications of this view when he says:

"I often ask patients to describe anything that they have observed or noticed about me that may shed light on aspects of our relationship...I find that it is critical for me to ask the question with the genuine belief that I may find out something about myself that I did not previously recognize... in particular, I focus on what patients have noticed about my internal conflicts." (p. 37)

Investigators like Aron and Hoffman are mindful of the fact that an analyst cannot participate anonymously in the clinical situation. They recognize that a pretense of anonymity cultivates idealization of the analyst as authority, and are most concerned to be sure that expressions of the analyst's subjectivity become matters for discussion as treatment unfolds. However, the technical approach that these authors advocate still implicitly assumes the possibility of at least relative anonymity for the analyst: disclosure of the patient's perception of the analyst's subjectivity is invited, but explicit communication of the analyst's perception of his or her own subjectivity is not equally recommended. A stance of anonymity is not entirely relinquished, even as the myth of the analyst's anonymity is analyzed. Greenberg (1991), for example, offers the following rationale: "...self-revelation can foreclose full exploration of the patient's observations and his reactions to them. My technical prescription... is not to confess but to follow the often more difficult path of maintaining an awareness of the plausibility of the patient's perceptions." (p. 70). However, I believe Hoffman's (1994) candid admission goes to the heart of the matter: "The magical aspect of the analyst's authority is enhanced by his or her... anonymity. There is a kind of mystique about the analyst that I doubt we want to dispel completely." (p. 198). We may not want to dispel it, but I think we should!

It seems to me that if we look at the work of analytic thinkers who are trying to develop a theory of technique that takes into account the truly intersubjective nature of the psychoanalytic enterprise, one that does not cultivate idealization of the analyst, we see general recognition of the need to move beyond our traditional ideas about self-disclosure. However, a systematic conception that can replace the principle of analytic anonymity has not yet been worked out. On the question of self-disclosure, even very innovative thinkers tend not to go beyond open-ended, non-specific formulations. Aron (1991) says: "The question of the degree and nature of the analyst's deliberate self-revelation is left open to be resolved within the context of each unique psychoanalytic situation." (p. 43). Ehrenberg (1984) summarizes: "Too much of one's... participation can destroy the integrity of the analytic relationship, as does too much caution. What is obviously needed is a delicate, judicious balance which establishes optimal distance." (p. 565). Burke (1992) warns against an "unwavering position" one way or the other on the issue of "countertransference disclosure," and suggests maintenance of a balance between "asymmetry" and "mutuality." We can certainly agree with such general position statements, but they do not offer us very much direction with respect to everyday practical clinical choices about what to tell our patients.

Clearly, some forms of self-disclosure by an analyst can be helpful and others harmful. I think we have ample reason to conclude that the categorical principle of analytic anonymity -- i.e. that, all other things being equal, communication to the patient of personal information about the analyst hinders analytic investigation -- is not valid; but what more can we do to develop useful criteria concerning self-disclosure?

Taking into account all that I have noted up to this point, I would suggest that we need to begin by not just discarding the principle of analytic anonymity, but by contradicting it: I propose that it is useful for the analyst consistently to try to make sure that his or her analytic activity is understood as fully as possible by the patient. I think it is best for an analyst to present a patient with a clear and explicit picture of the analyst's conscious view of his or her purposes and methods. An analyst should aim for comprehensibility, not inscrutability. I am not advocating imposing one's thinking upon a patient, but I am suggesting that one's thinking should be made available. For instance, if an analyst's intention in making an intervention is not self-evident, it can be a good idea for the analyst to make it so; if the understanding that informs an analyst's conduct is not obvious, or if the evidence (as perceived by the analyst) for that understanding is not obvious, it is usually helpful for the analyst to explain them. This is not to say that an analyst always has a clear idea in mind of what he or she is trying to do. Sometimes the analyst's perplexity, or the spontaneous, un-thought-out nature of a remark the analyst has made, is what needs to be stated explicitly for the patient to consider.

Now, my impression is that most of us have been taught to do the very opposite of what I am recommending. We have been encouraged to keep our intentions and assumptions to ourselves, to avoid explaining our activities to our patients. As a result, we tend to be ambiguous rather than anonymous. By declining to disclose what he or she has in mind, an analyst does not become a blank screen, or a mirror, or even a Rorschach blot. Rather, by acting without explanation, the analyst essentially poses a riddle. The analyst's behavior could signify a number of things, but the patient doesn't know what the analyst meant. The patient is asked to select from among multiple choices, one of which is favored by the analyst without the patient's knowledge.

Not knowing the analyst's construction of reality does not help a patient identify and reflect upon his or hers. On the contrary, it interferes and distracts by implicitly inviting the patient to guess what is in the analyst's mind. Whereas an analyst's effort to be anonymous is supposed to allow the patient greater freedom to associate, the opposite is the case, in my experience. An analyst's preferences, inevitably communicated in the analytic situation, are all the more influential and inaccessible to review for being inexplicit. By pretending to anonymity, an analyst increases the constraint he or she exercises. Far from diminishing the analyst's presence, a stance of non-self-disclosure tends to place the analyst center stage. It makes the analyst into a mystery, and paves the way for regarding the analyst as an omniscient sphinx whose ways cannot be known and whose authority, therefore, cannot be questioned.

On the other hand, when an analyst tries to communicate his or her thinking in full, respect for the patient as collaborator is conveyed. By publicly (within the treatment) taking responsibility for his or her own psychic reality, an analyst both requires and allows opportunity for a patient to do the same. Of course, explanation by an analyst of how the analyst sees his or her analytic activity is no guarantee against idealization by the patient. Obviously, it is at least as easy for an analyst to be idealized for being open, candid or iconoclastic as for any other reason. We have only to remember the old Jewish man who gazed at himself in the mirror and mused, "You know, I'm not very good- looking; and I'm not very smart; and I'm not very rich; but boy, am I humble!". The point of an analyst presenting the analyst's own view of his or her work as a subject for discussion is not that this prevents the analyst from being idealized by the patient, or even discourages it, but rather that self-disclosure of this sort makes the analyst's way of operating, like the patient's, a legitimate subject of joint inquiry. Thus, identification and correction of unproductive technique -- including the analyst's wish to be idealized, if that is a factor -- is facilitated.

For me, the what and how of self-disclosure consists of the analyst trying to communicate what are in the philosophical tradition termed pense‚s pense‚s, that is to say, the analyst's thoughts as they have been thought. I try to make my understanding of my participation in our work together as available to the patient as I can. When an analyst intervenes, it is because the analyst feels he or she has something to say that may contribute to the patient's self-investigation; therefore, I understand the logic of self-disclosure to be that an analyst tries to communicate any thoughts that are pertinent to the potential contribution, as the analyst sees it.

Often, I believe, discussions of the problem of self-disclo-sure are inadvertently skewed by casting the question too narrowly, in terms of whether the analyst should report his or her emotional reactions: Bollas (1987) refers to "the expressive use of countertransference," Burke (1992) to "countertransference disclosure," Ehrenberg (1984) to "affective participation," and so on. Since in fact an analyst's feelings and intimate responses are expressed in everything the analyst says and does in the clinical situation, I don't think it makes sense to equate the problem of self-disclosure with an analyst's decision to reveal "affect" or "countertransference," as if these categories denoted distinct and isolable aspects of the analyst's mental life. It makes no sense to consider any one category of personal information about the analyst (feelings, judgments, values, opinions) in and of itself problematic. Instead, it seems to me that an analyst's decision concerns which of his or her thoughts -- always an inseparable amalgam of cognition and affect -- to articulate; and I would say that an analyst should try to articulate and communicate everything that, in the analyst's view, will help the patient understand where the analyst thinks he or she is coming from and trying to go with the patient.

I emphasize in the analyst's view because, clearly, patient and analyst may disagree about what it is useful for the analyst to disclose, in which case the matter becomes open for consideration -- neither the analyst's nor the patient's view being privileged a priori. For example, a patient has the idea that I was being extremely gentle and careful with him the prior hour because I was afraid of hurting him. I respond that I was not aware of any particular concern on my part, and that therefore, from my point of view at least, the patient has his own reasons for imagining that I consider him so fragile. I feel it is useful to make explicit my own perception of my emotional state during the hour in question, since it is partly upon that perception that I base my hypothesis that the patient has an ulterior motive for experiencing me as gentle and cautious.

However, at a different moment in the treatment, the same patient feels, when I point out to him that he didn't seem to recognize that a remark made by his girlfriend was very disdainful, that I am trying to influence him to break up with her. He believes I have reached a judgment, based on what he has told me, that he can do better than this woman; and he wants to know from me what impression I have formed about his relationship. I answer that whatever private opinions I may or may not be entertaining are not really relevant to our purposes: he knows very well that his need to deny his girlfriend's disdain is an important matter, and doesn't really indicate in itself whether the relationship is worthwhile. In any event, since she's not my girlfriend, it's not my opinion of her that counts. Therefore, I see his preoccupation with getting a judgment from me as an avoidance of exploring the purposes of his denial of his girlfriend's disdain, as well as a wish to have me make his decisions for him.

In each of these two instances, I reached a different conclusion about self-disclosure, based on my view of which of my thoughts were relevant to what my patient and I were collaboratively trying to understand about his mental life. In the first instance, I decided to state a perception of my own; and in the second instance, I decided not to state one.

The problem of self-disclosure by the analyst is sometimes discussed in terms of whether the analytic relationship is mutual or asymmetrical (Burke, 1992; Hoffer, 1992). In my view, the psychoanalytic situation is one of what I would call complete epistemological symmetry: that is to say, analyst and analysand are equally subjective, and both are responsible for full disclosure of their thinking, as they see it relevant to the reality of the psychoanalytic endeavor. We might use as a motto for the analytic relationship a remark attributed to the filmmaker Frederico Fellini: "The only true realist is a visionary, because he testifies to his own reality."

However, symmetry is not identity. The thoughts of analyst and patient are very differently organized because analyst and patient have different functions in the clinical setting; each is oriented to their shared endeavor from a different vantage point. Whatever immediate purposes may come into play, ultimately a patient communicates his or her own reality in order to increase his or her own self-awareness, whereas an analyst communicates his or her own reality in order to increase the self-awareness of the other person. Form follows function, which is why self-disclosure for a patient consists of an effort to free associate, whereas self-disclosure for an analyst is deliberately selective. The difference between the self-disclosure of the analyst and the self- disclosure of the patient is not how much, but according to what principle. Ferenczi's much-criticized experiment in "mutual analysis" (in which he and the patient took turns upon the couch saying whatever came to mind) went astray not because Ferenczi's self-disclosure was excessive, but because his self-disclosure was organized in relation to a misguided objective. Ferenczi tried to accomplish the simultaneous analysis of two individuals within a single analytic setting -- an over-ambitious effort that was doomed to failure.

Even when the goal remains analysis of the patient alone, there are many possibilities for how useful self-disclosure is accomplished by an analyst. Different analysts have different levels of ease with exhibitionism. Some talk readily about themselves, others are more reserved; some don't mind meeting with patients face-to-face, others cannot stand being looked at; some are able to lecture when their patients are in the audience, others find this situation constraining. What is important, it seems to me, is that an analyst's personal preferences be dealt with candidly for what they are, not imposed upon the patient in the guise of analytic technique (and not, in the same guise, imposed upon students and colleagues as a categorical imperative).

Certainly, it can be to the patient's benefit that an analyst establishes comfortable working conditions for himself or herself; on the other hand, an analyst's comfort can be obtained at the patient's expense. Relinquishing a stance predicated on the pretense of anonymity deprives the analyst of protection from a kind of explicit, unameliorated scrutiny that can be most distressing; but we are obliged not to take refuge under cover of technique. I think the great majority of successful clinical analyses require that at certain points, the analyst, like the patient, accept the necessity to depart from his or her own preferred ways of proceeding and to bear a measure of discomfort.

By acknowledging that an analyst's judgments concerning what constitutes relevant full disclosure on his or her part are subjective, we indicate a role for the patient as constructive critic of those judgments. This is the reciprocal of the analyst's familiar role as critic of the patient's self-disclosure. We know that when a patient tries to say everything that comes to mind, an analyst is able to point out things the patient overlooks. Similarly, when an analyst tries to make his or her analytic activity as comprehensible as possible, a patient is able to point out things the analyst overlooks. I think Mitchell (1994) describes the analyst's position in relation to self-disclosure cogently when he says, "I am not necessarily in a privileged position to know, much less to reveal, everything that I think and feel" (p. 9).

A patient ends an hour one day by complaining that he senses I am not happy with him. He thinks there is a slight irritation in my tone. He's been talking all hour about a painful rejection, and now, on top of it, he feels rejected by me too. He leaves and I reflect. He's right. I am a bit exasperated with him. Why? In the treatment recently, we've been talking about a way that I believe he shoots himself in the foot socially. He wonders why he doesn't have more friends. I've been suggesting that the same unconscious competitive strivings that used to cause him so much trouble in his romantic relationships (he has made great gains in that area and is very thankful) are still getting in his way with regard to potential friendships.

This hour my patient has been describing how his colleague has struck up a relationship with the boss. My patient is unable to tolerate his envy, and it's paralyzing him at work. He feels so hurt he can barely talk to his colleague, his boss, or anyone else. He can't accept the situation, and I'm trying to help him understand why. He has a need to be #1 that bears looking into; but when I try to address it, he doesn't seem to understand what I'm talking about. He keeps berating the boss for rejecting him and himself for screwing up. "Why doesn't he like me? What's wrong with me?" he asks bitterly. His assumption is that he ought to come up on top, and if he hasn't, it's because he's made a mistake; yet, he doesn't see that this is a very rivalrous attitude, let alone that it may have something to do with why his boss, among others, doesn't take a shine to him. Instead, his view is that his boss rejects him for mysterious reasons; and I'm doing it too.

He begins his next hour by re-stating his complaint about the prior session. I answer that, indeed, I had felt put out with him because he was thwarting my efforts. I admit that this was an uncalled-for, self-centered reaction on my part. I say further that at the same time, as I reflect on how and why he got under my skin, it seems to me that his way of relating might well have provoked even someone nicer than me: when I tried to invite him to look beyond his self-pity, he ignored what I was saying, then put me down for being unsympathetic. I acknowledge that another analyst might well not have been provoked and might have been able to retain a friendlier attitude; that was my problem. His contribution was an attack on me of which he was apparently unaware. We discuss the preceding hour. He begins to be able to see that he has a tendency to attack anyone who causes him to experience envy, often feeling like a victim all the while he is insisting on his right to be #1; and in the hours that follow, he realizes that what went on between us is an example of exactly the kind of thing that obstructs his friendships.

My patient's comment prompted me to fuller awareness and disclosure of my view of him, which opened a way for us to fruitfully investigate his concerns about the reliability of my helpfulness, his characteristic, maladaptive ways of managing envy and competition, and a variety of other important, related factors. Now, the initial irritation in my tone had been quite mild, in my judgment, and I thought my patient's perception of it was strongly colored by his expectations. I could easily have treated my patient's complaint about me as plausible and remained non-committal myself. I could have asked him to elaborate and explore his ideas about my state of mind. This more traditional approach might have worked out very well. It could also have allowed us to stay bogged down in speculations from my patient about my attitude, leaving him continuing to feel rejected without being aware of his tendency to provoke rejection. In any case, I don't think any opportunity was lost when I followed the course I did. I took responsibility for my view of how I had both attacked and been attacked, which had the beneficial effect of requiring and helping my patient to do the same. It permitted my patient to reflect on his experience of me as an authority, rather than to continue to live it out within the treatment relationship.

I'm very much aware of the difficulty of effectively describing how my view of self-disclosure translates into action, let alone what I think are the advantages of my view of self-disclosure. Whenever one offers a clinical vignette intended to show the utility of a technical innovation, one is open to the comment, "that's okay, but it would have been better if you'd done it the usual way" -- a criticism that can never be disproved, since controlled testing is not possible. Clearly, anecdotal case examples don't constitute evidence; they are merely illustrative. Even as illustrations they are open to challenge. We often hear said about a specific application of a general technical principle, "I do that sort of thing all the time anyway. Why do we need the conceptual revision you are proposing?". The purpose of improving our theory is to make it something we don't have to either be unproductively constrained by or ignore. We want theory to be a tool that helps us find our way to successful technique more of the time. It is with that goal in mind that I offer a revised view of what I understand to be the traditional ideal of the anonymous analyst. Having stated these caveats, I'd like to mention a few more ways in which my approach to the problem of self-disclosure directs my clinical activity.

Because I am less hesitant to make my thinking known explicitly and in detail than I was some years ago, I find that now I'm more likely to share with patients certain questions that formerly I would have felt I had to decide for myself as matters of technique. For example, a young man is going over at great length the details of a decision he faces at work. It's my impression that he has passed the point of constructive thinking and may now be engaged in a process of rumination that has a motive not immediately evident to us. I'm considering addressing this possibility in some way. At the same time, I know that this young man is extremely sensitive to criticism -- throughout his childhood, he experienced his father as dissatisfied with him, finding fault with everything his son did. If I question the purpose of my patient's current thought processes in any way, no matter how diplomatically and respectfully, it is very probable that he will feel put down. This reaction on his part will then of necessity become the phenomenon of interest to us. I can intervene in the way I'm considering, and, if things turn out as I anticipate, investigate the patient's ideas about my disapproval -- perhaps externalizations and projections of his own misgivings about his rumination will eventually be unveiled; or, I can hold off, waiting to see whether he comes to reflect on his own about the motivations for his apparent rumination. Earlier in my career, I would have regarded this situation as consisting of a technical choice that I, the analyst, had to make. Now I would more likely share the dilemma, laying things out to the patient very much as I have just described them, including my speculations, my objectives and my concerns.

I think this kind of self-disclosure by an analyst is basic to an attitude in which the clinical enterprise is conceived of as a true collaboration between peers. In my view, whenever an analyst keeps his or her objectives, methods, or assumptions private, it privileges the analyst's point of view and maintains an idealized image of the analyst as superior to the patient. (This is true even of some of our most benign and humane conceptions of the analytic relationship, e.g., Loewald's [1986] notion of a "gradient," along which the analyst's relative maturity pulls the treatment forward.) Faced with a clinical dilemma, an analyst should feel at least as ready to seek consultation from the patient as from a colleague. Sharing the dilemma between analyst and patient explicitly acknowledges the true state of affairs, which is that each analytic couple has to negotiate its own way of working (see Pizer, 1992). Obviously, we cannot transcend the problem of the analyst establishing himself or herself as an authority, since it is inherent in any decision an analyst makes, including decisions about what he or she will disclose; but we can acknowledge the problem and begin to establish a mechanism for self-correction by inviting our patients to join us as collaborators, even in questioning our methods (including our decisions about self-disclosure).

Because of my view of the utility of analytic self-disclosure, I feel freer than I might otherwise to communicate certain perceptions of my own to patients. For instance, Adler (1994) talks about how, with borderline patients, it can sometimes be very helpful for an analyst to say when he or she thinks treatment has been going well, thus contradicting and identifying as possibly symptomatic a patient's apparently irrational negative evaluation. In my own experience, the sort of situation Adler describes comes up with all kinds of patients; and what Adler conceptualizes as a technical modification necessitated by a patient's unusually severe psychopathology, I regard as consistent with the principles of ordinary analytic activity. I not uncommonly offer impressions, optimistic or skeptical (see Renik, 1994), about the progress of treatment when I think they are apropos. For me, the crucial issue is that the analyst's judgments about analytic events be treated as subjective rather than authoritative. If I'm more positive than my patient about what our work together has achieved, I may have self-serving reasons for being so. If that possibility doesn't come up for consideration, we have to wonder why. On the other hand, if I question whether an analysis is going anywhere and my patient doesn't, we ought to at least pay some attention to the idea that a personal frustration of my own might be coloring my judgment.

All in all, I find that self-disclosure for purposes of self- explanation facilitates the analysis of transference by establishing an atmosphere of authentic candor. When my patients experience me as saying what I really think -- about them, myself, us -- they respond in kind. All too often, it seems to me, clinical analysis deteriorates into a game in which the patient feels free to bring up all sorts of ideas, without taking any of them quite seriously. When the analyst doesn't disclose what he or she is really thinking, and disclose it as completely as straightforwardly as possible, the patient is not encouraged to do so either. Disavowal gets built into the analytic discourse from both sides, and the patient's exploration of his or her experience is vitiated by a speculative, hypothetical, "as-if" quality. My experience is that the hardest thing for a patient to do is to discuss with his or her analyst profound convictions about the analyst's real character, to tell the analyst the sort of things that the patient suspects the analyst probably hears from friends and family members. Often, it is only in a second analysis that a patient feels free to consider what he or she really thought about a first analyst.

Of course, underlying my thinking about technique is an assumption about the mechanism of action of clinical psychoanalysis: namely, that therapeutic benefits are most extensive and enduring when they are based upon expansion of the patient's self-awareness. Certainly, a great many unresolved questions remain concerning the role of "insight" in a psychoanalytic "cure"; and an analyst's decisions about how to manage self-disclosure will necessarily be informed by his or her particular theory of how analysis works. However, to the extent we can agree that when possible it is best for therapeutic experiences to be consciously examined, the analytic aim of analyzing rather than cultivating unearned authority for the analyst cuts across an array of theories of the analytic process; and therefore it seems to me that an ethic of candor, implemented via self-explanation, applies as an overarching technical attitude (alongside whatever other criteria an individual analyst brings to bear when making decisions about self-disclosure).

We are understandably slow to question our basic assumptions. Being therapeutic practitioners, we are obliged to be careful about changing something that seems to be working well enough. Certainly, a great many successful clinical analyses are conducted by analysts who try to avoid self-disclosure in pursuit of what they think of as a stance of anonymity. On the other hand, we can't assume that everything that happens in a successful treatment contributes to its success. An analyst can be effective when elements of his or her technique are inconsequential, or even counter-productive. It is also true that a great many analyses conducted from a stance of anonymity and non-self-disclosure by the analyst founder or become protracted and unproductive; and a great many patients are deemed unsuitable for analysis conducted along such lines.

As I see it, we are now at a point where the evolution of our understanding of the epistemology of the analytic situation requires us to discard the ideal of the anonymous analyst, and we are left with the problem of how to systematically characterize the most useful way to present our thinking to our patients. I have suggested that as a first step we need to re-define self-disclosure. I've tried to describe my own thinking in this regard, and I hope the medium has been the message!

Abend, S. (1986), Countertransference, empathy and the analytic

Grune and Stratton.

---

---

Last Updated: February 13, 1996