Excerpts from Must Read Books & Articles on Mental Health Topics
Articles- Part X

MORTAL GIFTS: A Two-Part Essay on the Therapist's Mortality
Journal of The American Academy of Psychoanalysis, 30(2), 173-204, 2002
Ellen Pinsky

Part 1: Untimely Loss
How should this grief be properly put into words?
Melpomene, to whom the Father gave
The voice that to the music of the lyre
Flows out in mournful measure, teach me the art.
-Horace, from "To Virgil," translated by David Ferry 

Abstract: The absence of theoretical and practical provisions for the patient whose therapist dies or becomes seriously ill reflects underlying problems regarding termination. Therapy is unique among human intimacies in that its goal is separation, a paradox that informs both the near-silence of early termination theory and the confusion in more recent writing. The therapist's emotional involvement must be understood through the therapeutic relationship as part of "mortal" life, that is, as a specialized category within ordinary human interactions. The profession has neglected the therapist's mortality, in figurative as well as literal senses. This neglect, a covert grandiosity, is the "Olympian Delusion." On one level, inadequate termination theory underlies failure to confront the therapist's mortality; more profoundly, failure to confront the therapist's mortality underlies deficiencies in termination theory. The mystique of the superhuman therapist can lead to a professional reticence that is less than fully human, abrogating the patient's right to a decent, human leavetaking.
Key Words: illness; termination; mortality; death; therapeutic relationship; mourning; loss

 When my therapist suddenly died at the end of May in 1994, I felt lonelier than ever in my life. 1 turned for comfort to the same sources I have always turned: to people and to books. The first were wholly satisfactory, the second only partly so. In turning to people, there were two directions to go, the personal and the professional. Lucky in the safety net provided by the comfort and understanding of my friends and family, I was additionally fortunate in the strong support tendered by my therapist's professional community. His colleagues offered themselves immediately and generously, at every point proceeding carefully and wisely, all in the face of their own confusion and grief. They remained available throughout the immediate crisis, making clear their availability if needed into the future as well. In my view, this full assumption of responsibility for the patient's well-being provides a model. I was truly fortunate in people, in both the personal and professional realms.
    However, I can't say the same for the comfort I got from reading. In turning to books, I also had two accustomed directions to take: first, I could go to literature, which has cushioned and buoyed me since I was small, and, second, I could go to psychoanalytic theory. In literature I found what I needed; but in psychoanalytic theory, not only did I not find what 1 needed, but what I did find outraged and discomfited me. Looking for a way to lessen my pain and confusion, I turned to the body of papers on the illness and death of the therapist. I found instead of comfort a great deal of avoidance, confusion, condescension, self-delusion, and dissimulation. Rarely did I find the voice of the patient.
    My purpose here is not to review these papers, something others have done well elsewhere; rather, I aim to illuminate some factors that underlie the manifest difficulty the profession has had creating a literature adequate to a subject touching our deepest anxieties: we all die, we are all at every moment vulnerable, and we are all inclined to deny the fundamental terror. I raise far more questions than I answer. In taking a strongly critical stance about the profession's failure to adequately address the therapist's mortality, I do not mean to under­value the courage required of every author who risks approaching this difficult subject. Most especially, I hope my admiration is apparent for those whose search for understanding in the frightening, painful but vitally important matter of human vulnerability is particularly respectful and truthful. 

COMMONPLACES

People really die, not only some but all, each of us in our turn. --Freud, from "Thoughts for the Times on War and Death"

When people part, whether for the moment, for a while, or forever, they will say good-bye. This ritual of parting holds whoever the people are, whether greengrocer and customer, teacher and pupil, lover and beloved, two brothers, or two friends. The more important they are to each other, and the longer the anticipated separation, the more important it is to take leave well. These are just commonplaces.
    In matters so thoroughly ordinary, we may assume that the needs of therapist and patient are no different from the needs of other ordinary people; and it follows that the greater the depth of the therapeutic relationship, the greater the need at parting for an adequate good-bye. These simple truths hold, whatever the otherwise peculiar qualities of the psychotherapy situation. In the odd technical language of psychoanalysis, this work of parting is named "termination," and it is the endpoint toward which every therapeutic engagement naturally moves; that is because, unlike other relationships of depth and intimacy, the therapy relationship intends right from the outset to end. Here we are faced with a paradox: the goal of this special form of closeness is separation.
   In his cogent examination of the word, Pedder (1988) suggests that termination "is a curiously inappropriate term with its negative and finite connotations which fail to convey the positive hopes for a new beginning that normally surround the end of a satisfactory analysis" (p. 504). However, his suggestion that the term is exactly appropriate when the ending is forced or premature is strikingly germane to the topic under consideration, the threatened and sometimes literal "termination" of the analyst's life. But we might then wonder what lies behind the adoption of such a "curiously inappropriate" word to describe the natural ending of the analytic relationship. Perhaps the relationship's paradoxical intent--an intimacy that intends separation--builds into every analysis an inherently traumatic ending, and in that sense the word "termination" is quite apt, capturing an emotional truth in the experience of parting.
    Ekstein's (1965) distinction between positive trauma and negative trauma may be useful here; referring to the Freudian model of thought in which delay and frustration explain both "the development of psychopathology and the development of creative and healthy function" (p. 63), Ekstein points out the need in such a model "to differentiate between the positive trauma, which leads to creative solutions, and the negative trauma, which leads to the experience of being overwhelmed and to pathology" (p. 63). Ekstein likens termination to the epilogue of a play; emphasizing successful mastery as central to the process of ending a therapy, that epilogue "now becomes the prologue for another play”:

The epilogue, the termination phase of the analysis, sums up once more what has taken place. It is a kind of Thanksgiving, a good-bye and a mourning, and a cautious trying out of new wings. (p. 680

 Termination becomes the final adaptive act of mastery without which all that precedes it is not complete; in these terms, it is a "positive trauma" and a creative act.
    But while the work of a sound ending necessarily includes the painful task of acknowledging the human limitations of each partner, and likely the limits of therapy itself, the therapist's literal mortality, the fundamental shared human limitation, is not supposed to precipitate that ending. (Although it is not my subject here, it should be said that neither is the patient supposed to die.) As Winnicott's 12th rule states, "The analyst survives," and anything else is a breach of expectation. And the more importance theory gives to this ritual of ending--that is, to the proper "termination" of a therapy--the greater, in theory, the difficulty for the patient who loses the partner. In Ekstein's terms, there is potential in such an ending for "negative trauma."
    Every patient, in embarking on a therapy, risks attachment, dependence, and intimacy, quite possibly without knowing what he or she risks; after all, the patient can't know at the start the power of the transferential relationship whose sweep "has the dynamic potential to reawaken every past desire for communication and love, protection and admiration, that the patient has experienced in the course of his life. Together with the desire and the hope come the fears of past rejections and failures, traumas and life injuries" (Rizzuto, 1998, p. 5). Nor is the embarking patient likely to consider, much less be able to measure, his isolation in this most private of professional affairs. The patient expects the joint labor to bear fruit, with the therapist remaining to an agreed-on ending. The implicit promise of constancy makes risk tolerable: the helper will be there for safety and guidance through to that ending. One's connection to other professional figures--a lawyer, accountant, or cardiologist, for example--bears little comparison, as is true even of the connection to a beloved teacher or mentor, a priest or pastor, though these may provide a better analogy.
    It should go without saying that this summary does not mean to blame the therapist who becomes sick or dies for his human condition; the analyst, like the parent (and all others who assume roles of guardianship), knows that he "has implicitly promised what he cannot possibly guarantee" (Ogden, 1997, p. 11). Rather, because any therapeutic contract by its nature creates a situation which fosters emotional dependence on the professional caretaker, in every instance a mortal man or woman, it would seem an obvious responsibility of the profession, in an institutional and theoretical sense, to look hard at this contingency. It is, after all, a loss which moves onto center stage a new problem for the patient, one that has been imposed by the vehicle through which he has sought help. Without blame, there has been a catastrophic breach of the therapeutic contract itself.
    In this respect, however, the profession has been remarkably slow, even unconscionably slow, to take responsibility for the tremendous power of its very method. Gardner refers to the "failure of psychoanalysts to practice the self inquiry they preach" (1989, p. 6). If the therapist's job (as Winnicott says) is in some sense to survive, and if the therapeutic endeavor is a self inquiry requiring of the patient an honest, unwavering examination of the most frightening truths about himself, how shall we reconcile the fact that therapists get sick and die with the fact that, historically, remarkably little has been written either about such loss or the underlying realities it raises? "Death terrifies us all into silence," Margulies writes (1998, p. 3), the simplicity underscoring the profound truth; but responsibility is not thereby abrogated.
    In the scholarly writing that does exist, themes of omission and denial emerge repeatedly, an "affect filled silence" (Schwartz and Silver, 1990, p. 2) continuing to shroud the profoundly difficult subject even as the literature slowly grows. And when the subject is addressed, the chief focus is almost always on the therapist: on his dilemma, his perception of the patient's resultant dilemma, his decisions about what will work best for the patient, and his judgment about the effect of a chosen intervention. Silence about the matter of his own mortal vulnerability being a frequent choice; the patient's voice throughout is very faint. Ultimately, as in any human relationship, the dilemmas of patient and therapist are inextricable.
   Freud is reported to have liked quoting Charcot, "`Theory is good; but it doesn't prevent things from existing"' (1893, p. 13, n. 2). It also does not prevent things from happening. If' the discipline of psychotherapy locates at its heart an intense, intimate and private relationship with the reliable professional caretaker, should not its theory incorporate the fact of that caretaker's mortality--an ultimate unreliability? The therapist's death suspends the therapy in a movie freeze-frame, isolating and magnifying the nature of the very peculiar relationship: professional and private, fiduciary and intimate, limited and intense. This suspension reveals the absence of provision for the patient, both theoretical and practical, when the therapeutic frame collapses. How may we account for this absence?
   The theory of termination ought to inform the "affect-filled silence" surrounding the therapist's mortal vulnerability. The link between the theory of termination and the event of the therapist's untimely loss should be a vital one. Termination theory should tell us about the theoretical ideal: leave-taking at the natural conclusion of the therapeutic engagement. That leave-taking, or farewell, is a necessary ritual performed within the therapy frame. But a paradox or anomaly arises from the conflict between the theoretical ideal of termination, in the technical sense, and the brute reality when the therapist's life is threatened or literally terminated.
    I want to emphasize that "untimely loss" has broader meaning than loss through the therapist's death; while the neglect of that most extreme eventuality is my narrow focus, "untimely loss" has at every moment wider implications, just as "mortal nature" implies far more about the therapist's human limitation than that he can die. For example, the sick therapist may be untimely lost, though temporarily, if his illness makes him absent; or, he may be physically present and simultaneously lost if his illness makes him distracted and inattentive; to take the reasoning further, the therapist who commits an ethical violation may in some sense be "untimely lost," profoundly so, even to those patients he has treated ethically. Clearly the idea means to incorporate the caretaker's entire "human being," and my subject in its broadest sense is the implications for treatment of the therapist's humanness.
    It could perhaps further be argued that every ending, even the most unproblematic, involves an "untimely loss" because, as Bergmann puts it, "On a certain date, every analysis comes to an abrupt end" (1997, p. 137, my italics). I am in this way returned to the idea of trauma inherent in every termination, the parting inevitably involving, in some sense, an "untimely loss." In the best of endings, this would approximate Ekstein's positive trauma, that "final adaptive act of mastery which is the dress rehearsal for future adaptive behavior" (1965, p. 62); at the other end of the continuum would be catastrophic breach, whatever form that takes: the worst of endings.
    If we can understand what theory postulates about the course of the natural ending to a therapy, we may be able to understand better the dilemma for the patient and therapist when ending is untimely and neither the therapy nor the termination process can approach an ideal conclusion. The more emphasis theory places on the importance of a termination phase, the greater the calamity when it cannot take place; and therefore the greater the weak spot in our theory itself, insofar as it makes insufficient allowance for the reality. 1 am considering, then, the potential (and always implicit) conflict between the theoretical ideal of termination--a process of mourning--and the fact of mortality.

SILENCE AROUND ENDINGS

For almost 75 years, psychoanalysts had been unable to conceive of the idea of a terminal phase ...Novick, 1997, p. 146)

As soon as we examine the history of how psychoanalysis has thought about endings, the silence about the therapist's mortal vulnerability makes more sense: what we find is that, as far back as the inception of psychoanalysis and up to this day, there is a remarkable lack of clarity about how the psychoanalytic engagement concludes. Not only does termination theory enter the psychoanalytic discourse relatively late in the century, but there remains into the present a great deal of confusion about what comprises the course of a natural (unproblematic or ideal) ending. It is by now generally agreed that termination is an identifiable and important part of the therapeutic process (although just how important is still questioned by some), and few would any longer question that it involves the work of mourning; but beyond these two ideas there is little consensus. How may we understand this delay and confusion?
    Novick (1997) proposes that analysts have a deep-seated resistance even to thinking productively about termination, tracing the difficulty back to the very beginnings of psychoanalysis. As the first analyst, everything Freud did was an experiment. He had a great capacity to notice mistakes and acknowledge failures, using them to further refine his theory. "It is therefore significant that neither he nor his followers seemed able to look at termination practices with the same scientific objectivity and courage" (1997, p. 153).
    Freud's significant omission must have meaning as well as consequences. Confusion in the part of theory whose purpose is to elucidate the completion of the therapeutic process, and to guide in its navigation, is in itself a curious gap, but it must also have implications for understanding the confusion around endings that are problematic. If the theoretical frame is itself wanting, obviously there is less to guide difficult contingencies. The psychoanalytic situation is structured to set in motion a process of remarkable power; the spectacle of the sorcerer's apprentice comes to mind, that small magician frantically trying to control the powerful forces he has unleashed. Bergmann (1997) aptly uses this image in a paper dealing with the concepts of the transference neurosis and the resolution of the transference as they relate to the process of termination. He writes: "If psychoanalysts can foster a transference neurosis but are helpless at resolving the transference they themselves have created, they may well be in the position of the sorcerer's apprentice who can begin a process that he cannot bring to a conclusion" (p. 146).
    In 1913 Freud presents the metaphor of the chess game, saying that the opening and endgames of chess, unlike the middle, could be systematically elaborated; making the comparison to analysis, he lays out rules for the beginning but does not go on to deal with the ending (Novick, 1997, p. 145). In Freud's conception, analysis, like chess, "could be brought to a successful end (not a stalemate), but the criteria, characteristics and management of the ending were never discussed" (Blum, 1989, p. 275, my italics). Of significance as well for the fate of termination theory, and for the more specific topic of the mortal therapist, is that, even after dealing explicitly with object loss in Mourning and Melancholia (1917), Freud "did not refer [in his writing] to the loss of the analyst and the meaning of the loss of the analyst's direct participation in the analytic process" (Blum, 1989, p. 279). This additional omission helps explain the late entry of mourning into the theory of termination.
    Nor do Freud's immediate followers attend to a termination phase (Hurn, 1971; Novick, 1997); except for small steps in that direction by Ferenczi and Rank (1924) and by Ferenczi (1924), the literature is silent until 1950 and remains sparse for another twenty years. For example, in 1950 Annie Reich writes, "Astonishingly, the topic of termination of analysis has rarely been the subject of psychoanalytic investigation" (p. 179), and in 1966 Rangell remarks on "the relatively scant literature" (p. 141). In 1971, Hum says of termination that "precise and detailed definitions of the term do not exist in the literature" and there is "no paradigm of the terminal phase" (p. 332). Freud's 1937 essay "Analysis Terminable and Interminable" does not deal either with termination as a phase or with techniques to guide it; abandoning the myth of human perfectibility, Freud concerns himself in the essay with "the inherent limitations of psychoanalytic technique, the patient, and the analyst" (Novick, 1997, p. 145).
    In light of Freud's advancing age and his painful battle, by then fourteen years long, with cancer, this concern with limitation has a special poignancy. It is clear in Freud's correspondence that he preferred his friends not inquire about his health, a choice his biographer Schur attributes to a combination of strength of character and avoidance: "Freud not only wanted to spare others the necessity of asking about his condition; he himself did not want to be asked about it" (1972, p. 378). Whatever Freud's motivations, it must be wondered how his handling of his own lengthy illness may have subtly permeated expectations within the culture of psychoanalysis, both written and unwritten, for the failing analyst's conduct. Theory, whether articulated or implicit, may in such profoundly difficult matters ease anxiety, serving self justification and denial.
    At any rate, in 1939 Freud died, never having referred to the effects of his disabling cancer or impending death on his work with patients, this fact in itself a noisy commentary on the silence around the therapist's mortal vulnerability. Schwartz (1993) points out that even Jones and Schur, Freud's analytic biographers, make "only passing mention of the clinical impact of his obvious and incapacitating disability" (p. 143). Schur's "passing mention" refers to a letter to Marie Bonaparte "in which Freud apologized to her in a deeply moving way for having allowed his preoccupation with his cancer to keep him from recognizing a certain transference phenomenon in her analysis" (Schur, 1972, p. 382). As Schwartz suggests, "Perhaps clinical analysis as a science was still too young to document the seeming technical confusion introduced by real disease in the doctor" (1990, p. 143). My suggestion, carrying this idea further, is that the "seeming technical confusion" has a basis more profound than the youth of the science, and that it resides at the heart of the paradoxical endeavor of clinical analysis: to foster an intimacy that intends separation. The idea is consistent with Stone's conceptual iztion of the analyst as representing the mother-of-separation in the primordial transference (1961, p. 77).
    In the last quarter of the century there has been an explosion of attention to termination, the literature so fast-growing that "the range of views concerning the conceptualization and management of this phase of treatment is now vast . . ." (Novick, 1982, p. 329). In light of a literature suddenly enormous in both size and range of views, it is all the more intriguing that, as Stephen Mitchell puts it as recently as 1997, "There is less useful literature on the termination of analysis than on any other major feature of the work" (p. 26). Also in 1997, in an essay titled "Termination: The Achilles Heel of Psychoanalytic Technique," Bergmann bluntly states that, in spite of a vast literature, "psychoanalysis, and particularly the literature on technique, has so far failed to offer a paradigm for termination" (p. 163). A half century of silence is followed by a half century of confounding attention.
    Perhaps more fundamentally at issue than either paucity or confusion, then, is the manifest difficulty in understanding the ending. This difficulty harks back to my introductory statement of purpose regarding the neglect of the therapist's mortality: I began by trying to illuminate some factors that underlie the profession's manifest difficulty in creating a literature adequate to a subject touching deep human anxieties: "We all die, we are all at every moment vulnerable, and we are all inclined to deny the fundamental terror." That formulation implies that weakness in termination theory underlies the failure to confront the therapist's mortality. However, the more I investigate this subject, the more I feel that the reverse may be closer to the truth: fear of mortality underlies and clouds the dilemma of termination.

THEORY: IT'S JUST SO MUCH CULTURAL DISCOURSE

Can it be that psychoanalysis, as Novick suggests, cannot conceive its endings and if so, how have therapists managed to guide treatments to their conclusion for a hundred years--as indeed, practically and experientially, we certainly have done? This gap between what we say we do and actual practice could thus be called benign, pragmatically. But to close the gap even a little can only benefit both theory and practice. In a tour de force presentation, Novick isolates five obstacles "that make termination inconceivable"; he examines factors "in the history of psychoanalysis, the training of the analyst, the counterreactions of analysts, the theoretical models held by analysts, and the kind of termination experience analysts themselves have had that create continuing obstacles to conceiving of termination" (1997, p. 147). I am in agreement with him and am also proposing a perhaps deeper difficulty.
    If there is consensus about the theory of termination, it is about inadequacy, a problem of neglect and evasion that applies equally, and I think not coincidentally, to the literature on the death of the therapist. Remarking that clinicians have obviously been "bringing analyses to some conclusion or other all along," Kaplan asks: "So how can we have a large literature on a practical subject that omits most of the actual practice the subject entails?" (1997, p. 175, my italics).
    Presumably, the majority of therapies play out reasonably successfully, unburdened by dire intrusions; and presumably there are also many instances of sick or dying therapists managing extraordinarily difficult endings with dignity, courage, and skill. That is, Kaplan's question about a literature that fails to reflect the actualities of clinical practice is equally applicable to the mortality of the therapist. In my instance, as noted in the introductory comments to this essay, that extraordinarily difficult ending was successfully facilitated at a remove, by the dead person's community. To me this seems a striking example of responsible clinical practice expertly carried out under potentially disastrous circumstances; my contention is that such expertise must have theoretical implications and underpinnings, though unarticulated. Although there is a growing literature on the illness and death of the therapist, it has so far brought us no closer to an adequate theory or to the provision of technical guidelines (or clear institutional supports, for that matter) to help therapist and patient in managing the painful reality; yet therapists' illnesses and deaths have been bringing therapies to some sort of conclusion for a century and of course will continue to do so.
    Kaplan proposes one explanation for the disjunction between theoretical formulation and actual clinical practice: "Too much of the technical literature of psychoanalysis," he writes, "has been simply so much cultural discourse" (p. 176), and one must be cautious in drawing conclusions about the actual world from any cultural discourse alone; it is a fallacy "that cultural data--literary works, religious tracts, and so forth--directly embody actual social experience and practice" (p. 176). Tongue-in-cheek, he wonders: "How much about the actual sexual practices of priests would you expect to learn by consulting church edicts and other such writings?" (p. 176).
    He is referring to the tendency for psychoanalytic theory--the formal documents or "cultural data" of psychoanalysis--to reflect idealizations rather than empirical observations, theory often being the production of "highly cultivated minds" trying, especially in the early years of psychoanalysis, to make legitimate an emerging discipline; as that discipline has developed (and ironically, at century's end, it once again struggles to legitimate itself), the tendency to idealization lingers in the body of literature. Kaplan pokes at psychoanalytic writing:

It has required a cunning capacity to read between the lines of our professional journals to avoid getting drawn into demoralizing comparisons of one's own actual practices with the hypothetical practices of the idealized Standard Psychoanalyst. Ours is the literature that has given us the exemplary analyst, who, upon being told excitedly by the patient that the latter's wife has given birth to twins, inquires, calmly, "What comes to mind about two?" (p. 176)

    This analyst from an earlier time, whose exemplariness results from his being thoroughly analyzed, evokes the dilemma for the analyst who, if he experiences strong and merely human feelings for the patient, must strive to meet an impossible ideal; the danger is that in such striving for the impossible, the analyst tends to hide or deny his very ordinary and real human responses, these feelings being a sign of dangerous and theoretically damned countertransference. Furthermore, this kind of idealized self-expectation, creaking under the strain of denial, more easily accommodates the self-delusion that he has succeeded; and the less humbly he tends to view his capacity to control what he feels, or always to know what he feels, the closer he is to believing he inhabits a superior realm. Of course the only such realm is the realm of the gods; and we all know that only the gods need not consider their mortal nature.
    I do not mean to imply that the dilemma of this hypothetical analyst from an earlier time is simply and easily eradicated by the trend toward acknowledgement and acceptance of feelings in the analyst; whatever may be the particular benefits (or for that matter, the real dangers) of shifts in theory, my more important point is the merely human tendency -- universal and ineradicable -- to self-deception. It may take many forms.
    As the "keeper of the analytic process" (Calef and Weinshel, 1980, p. 279) -- a role of guardianship -- the analyst strives to be truthful; that is, he monitors his own tendency to self-deception, engaging in a self­inquiry (Gardner, 1989) parallel to the one he guides and guards for the patient. One fundamental goal of that self-inquiry (for both analyst and patient) is to chip away as truthfully as possible at this self-deception. Truthfulness, one might say, is thus a central concern for the matter of termination: possibly it is the definition of the end, the defining point of the end--it is the termination point reached.
    It is important to emphasize that I am not talking about the acquisition of a definitive "truth" but about an increased self-awareness leading both to a greater freedom and to a greater sense of responsibility; at the end of a successful therapy the patient has learned to speak "truthfully enough," and also to understand that speaking absolutely truthfully is a never achievable ideal -- a vital understanding. This success involves commitment to the possibility while accepting the impossibility. One remains, however, committed to the approach, just as psychoanalysis remains committed to articulating its best understanding of what it does. Margulies puts this elegantly: "And so we attempt to observe what we do -- and who we are when we are doing what we do -- knowing that our words are never adequate to the task. We have no other choice but to try" (1998, p. 2).
    "A culture," writes Henry Louis Gates, "defines itself not by what it permits, among the near infinitude of human possibilities; it defines itself by what it doesn't permit" (1997, p. 122). Even after all this time, the vast literature on termination does not reflect what practitioners ­- ordinary mortals, one might say -- have actually learned to do, hour by clinical hour, working over the years. Nor have the many useful threads of a coherent theory that are already imbedded in the large literature been successfully pulled; rather, they seem to be lost in a proliferating tangle. Why is termination theory in particular so problematic to psychoanalysis? Why have psychoanalysts been unable to examine what clinicians over generations have learned to do when they end a therapy? And why have psychoanalysts been unable to extract from the body of written theory what makes sense about endings, blending practice and literature to come up with a more satisfying, unified theory and more useful clinical guidelines? Each of these questions applies not only to "good enough" terminations but equally to endings breached by the therapist's illness or death.
    I can think of no better way to summarize the aim of the first part of this essay than by returning to the question at its beginning, the question posed not in my words but those of the ancient poet in the epigraph: "How should this grief he properly put into words?" Embedded in this question is an explanation for Schwartz's "affect-filled silence" about the therapist's mortality as well as the suggestion of an answer that begins to fill it. The "termination" that designates the end of the therapy relationship is an extended process of mourning, a "long drawn-out leave taking" (Loewald, 1988, p. 258) that requires the patient's experience of and expression of sorrow or grief. It is through this experience and articulation that the patient takes leave well (perhaps in both senses); with an adequate ending, the therapeutic relationship -- an intimacy that intends separation -- terminates, and the therapy is complete.
    But there is a second person in the room, and that person also loses a partner. Though the therapist's role is different, the task is the same and it involves leave-taking -- unless, that is, the commonplaces about partings with which I began this essay do not apply to the therapist alone, uniquely exempt. How any individual analyst manages his heightened experience during this farewell, and how he (or she) determines what he will articulate or will hold private in the course of the ending, is a matter of his temperament and clinical judgement, and is not my point here. Rather, my suggestion is that the difficulty psychoanalysis has in articulating its theory of endings, both good enough endings and disastrous ones, reflects a difficulty with the experience of loss and grief. That very human difficulty, I further propose, is located just as much, and sometimes more, in the psychoanalysts (and in the cultural documents of psychoanalysis) as it is in patients.
    And here there is a quandary: to perform the ritual of saying good­bye is not only the patient's task but also a right; it is part of the provision implicit in the therapeutic contract. Therefore, it stands to reason that if any individual analyst suffers to a high degree with this difficulty, and if the "documents" of psychoanalysis as a whole tend to reflect such considerable difficulty, then the grieving patient's task may be immeasurably harder. The helpers in this case become the obstacle.

Part II: The Olympian Delusion

Mr. L. remarked that you can't expect even psychoanalysts to be superhuman, especially when they are dying. (Freedman, 1990, p. 308)

 The ancient saying, De mortius nil nisi bene [sic] (of the dead speak nothing but good), is still an absolute command obeyed by everybody, surprisingly even by analysts. (Tauber, 1989, p. 179)

     When the god Zeus feels desire, he likes to come down from Mt. Olympus where the gods dwell and pleasure himself with mortal women. Often he transforms himself for the purpose of taking his satisfaction; the woman Europa, for example, perceives him in the shape of a bull "beautiful beyond all bulls that ever were, bright chestnut in color, with a silver circle on his brow and horns like the crescent of the young moon. He seemed so gentle as well as so lovely . . . " (Hamilton, 1969, p. 79). But the gods don't get involved. Once Zeus has ravished the woman, he goes back to Olympus where the inhabitants don't feel our merely human mortal misery and pain; the gods are apart: not only do they not feel or suffer consequences as we do, but they also live forever. Blinded into submission by the god's magnificence, gentle demeanor, and force, the woman must bear alone both the mortal burden and the child.
    Each of the two quotations above contains the same disturbing assumption: analysts, although still human, are closer to gods than other people; and when circumstances force us to face their ordinariness, we are surprised and disappointed. Mr. L. is a businessman whose analyst died during the therapy. What is disturbing in his remark resides in the words "even" and "especially." The imbedded proposition is that, if it were possible for anyone to be superhuman, then psychoanalysts would be among those most likely to inhabit that godly realm. But Mr. L. has learned that psychoanalysts cannot attain the level "superhuman," something he found out painfully through his analyst's behavior when that man was dying.
    The second epigraph reflects a similar delusion perhaps more bizarrely, again involving the word "even"; the writer is herself an analyst. What can it mean to state that "an absolute command," one that is obeyed by all of humanity, is obeyed as well by psychoanalysts--but in their case it is a surprise? I am calling this view of the therapist "The Olympian Delusion."
    These epigraphs together capture a dilemma. There is a tendency for the profession, as well as the patient (and, more problematically, the patient's partner), to idealize and elevate the therapist, seeing him or her as someone not subject, in the same way as other people, to vulnerabilities of body and mind--a definition of the therapist as apart, "god-like." The roots of the dilemma are at least double, both theoretical and experiential, fostered not only in the theory that conceptualizes the therapeutic situation, but in the experience of each participant in the actual clinical exchange as well.
    In the instance of theory first, definitions of the psychoanalytic situation and of the therapist's function tend to emphasize the uniqueness of that encounter and role--it is no ordinary exchange. Throughout the literature the analytic relationship is conceived as like no other, a notion articulated by Freud in 1915 in specific reference to the analyst's handling of the patient's transference love: the doctor's role "is one for which there is no model in real life" (p. 166, my italics). To give shape to the peerless therapist, then, one must compare that paragon to something that can be known, and that is exactly what Freud and his followers did, and continue to do, as analyst and analytic interaction are conceived and reconceived, for a hundred years now--from a simile of analyst as blank screen upon which fantasies are projected, to a parental figure holding a vision in safe-keeping (Loewald, 1988), to a participant in a mutual self-inquiry (Gardner, 1989). While it may be that the therapeutic situation both resembles other human connection and at the same time is not exactly like any other, the same can be said of many other relationships. But we don't find education experts (for example) regularly reminding us of the uniqueness of the student-teacher bond, nor is there much controversy about defining the essential role of the teacher; yet that pedagogical bond is also unique, the teacher, like the therapist, in a position of sacred trust.
    In the discomfort revealed by such reminding, repeating, and redefining we hear an echo of the dilemma elevating the analyst's role, so difficult to define, to a special realm, apart--a position dangerously close to "god-like." While shifts in theory may of course represent the effort to grapple with this very dilemma (as we see, for example, the analyst removed from a position of "god-like" authority with the gradual democratization of the relationship in the shift toward an intersubjective understanding of the situation), the dilemma is not thereby erased. In her commentary on Amy Morrison's (1997) moving paper describing her work with patients over the ten years of her illness, Chused (1997) addresses the potential for self-deception in adhering blindly to psychoanalytic concepts, no one of which is absolute; that caution holds for the time-honored ideas and equally for the newly fashioned: for example, she writes, "though analytic anonymity can be deceptive (to the therapist and to the patient), self-disclosure, in its motivation and its effect, can also be deceptive. One truthful self-revelation can be used to hide another, perhaps more important one--as all lawyers know" (p. 246). Paradoxically, in the effort to define and redefine the analyst's function, we may find both the dilemma and its best safeguard; that is, the role is hard to conceive and like no other, which sets the analyst apart, and at the same time it is through this on-going self-examination and redefinition that self-deception may be best averted.
    The second root of the dilemma that elevates the therapist is experiential and, naturally, entwined with the first. The analyst in the specialized analytic situation may take on what Fliess calls a "rare and exalted perfection" (1942, p. 225) as well as being a human being. For the patient the tendency to elevate the therapist to the realm of invulnerability is understandable, even invited, given the structure and specialized purpose of the situation, the task for each participant, and the place of illusion if their shared work is to get done; that same tendency, if it lulls the therapist, is of course less benign.
    Hoffman poses the question, "Could it not be argued that the whole ritual of psychoanalysis is designed, in part, to cultivate and protect a certain aura or mystique that accompanies the role of the analyst?" (1991, p. 83). It is an unusual role, even in mundane terms, evoked simply by Modell:

Where in everyday life can you find persons who, for an agreed-upon period of time, will place their own needs and desires to one side and be there only to listen to you and who are more than usually punctual and reliable and can, for the most part, be counted on not to retaliate and to he free of temper tantrums? (1991, p. 25)

Both partners assume that this unusual arrangement will continue for the benefit of the patient until a planned ending, an assumption which encourages in both people the fantasy of a therapist invulnerable to the ordinary contingencies of life. While that fantasy may be beneficial for the patient, it may prove hazardous if it blinds the therapist.  Modell also writes about the essential place of illusion in a therapy:

For both therapist and patient, the other person is experienced both as an individual in ordinary life and as someone transformed by the therapeutic process, a ... [transformation ] that we label as transference and counter­transference. Therapist and patient are also everyday people, but within the frame of the therapeutic process an illusion is created that can he described as another level of reality. (1991, p. 15)

But what kinds of beings are these who, within the therapeutic frame, inhabit "another level of reality"? They are at every moment in that habitation no less real, no less less ordinary "everyday people." While a break in the treatment frame or a rupture of the illusion is a breach of expectation for both people, the potential for such a breach is an ever present reality; after all, both people are in fact always mortal in body and fallible in character. But if the therapist, who guards the process, loses sight of the distinction between an idealization, an "as-if" that is a necessary, benign illusion and an idealization that is a dangerous delusion, then the therapist has adopted an Olympian stance.
    Another seed of the kind of thinking that elevates the therapist to superhuman can be traced to Freud's early optimism about the psychoanalytic method and its potential for cure, that is, for producing perfect health: "the myth of perfectibility" (Weinshel, 1990); in the counterpart of that myth, we may posit an analyst who is superman (or Mary Marvel) or magician: that is, one who has the capacity as healer to create mortal perfection. Far from disappearing, however, the myth of the fully analyzed analyst--the same myth implied in Mr. L.'s and Dr. Tauber's "even a psychoanalyst"--lives on in conceptions of the analyst with superhuman capacity to do this very hard work.
    Analysts continue, for example, to have extraordinary expectations for their own performance and that of their peers (Abend, 1986), that expectation also implicit in the tendency to deny imperfection, even the most gross imperfection, as in the example of sexual violation and a century spent denying it and hiding it. Sexual violation is arguably the most egregious and disturbing manifestation of the Olympian delusion. If expectations are so high for the practitioners of the profession, it follows that those expectations will extend to qualities required of trainees:

The Panel (1961) on selection of candidates for analytic training collectively provided such an array of aptitudes to be sought in the budding analyst that to one participant suggested a composite portrait of Pasteur and Thomas Mann. (McLaughlin, 1981, p. 651, my italics)

And that portrait is of the candidate before the training analysis!
    Having abandoned his hopes for the attainment of perfect health through psychoanalysis, Freud wrote in 1937, straightforward and modest in the claim, "Analysts are people who have learned to practice a particular art; alongside of this, they may be allowed to be human beings like anyone else" (p. 247). But what happens if the analyst is not permitted (or himself denies) his own mortal nature--that is, if he (or she) denies or suppresses realities like anxiety, anger, desire, temptation, irritability, fatigue, sadness, ill health and so on? Perhaps the greatest danger is that, in self-effacement--a form of self-idealization--there is easily a reversal wherein the therapy, an arrangement ostensibly set up to meet the needs of the patient, becomes a vehicle to serve the needs of the therapist. Once that reversal takes place, there is the potential, its harmfulness depending on the particular needs of the therapist and the vulnerabilities of the patient, that the therapy will be used, as Calef and Weinshel put it, "as a tool to be directed against others" (1980, p. 289). The result of self-denial is thus a paradoxical one: Obliteration of the self brings with it the even greater danger of obliterating the other.
    The analyst's role of course requires a certain denial of self, an act of generosity intended to remove the therapist's needs from obtruding, thus reserving the focus for the patient; self-effacement, it might seem, safeguards rather than endangers the patient. But that is not always the result. The more fully the therapist believes in his capacity for literally selfless service to the patient, and the more the therapist is conceptualized as above being a subject himself (or herself)--that is, as one who during the engagement is not subject to strong emotion, to sexual desire, to discouragement, to self-delusion in the very same ways that the patient is--the greater the danger of crossing the boundary that keeps the patient safe: a selflessness that blinds the therapist to his own need and vulnerability leaves him (or her) prey to the arrogance of perfection.
    Such a conception, in removing the therapist from the room as a subject, loses sight of the fundamental: There are always two people present. If there is only one subjectivity, there can be no boundary. The Olympian delusion is thus encouraged by a monolithic subjectivity. Since subjectivity can never be unilateral in reality, selflessness paradoxically inverts itself and becomes a tyranny. Both patient and therapist are then in danger.
    It is important to reiterate that the dilemma elevating the therapist, while it is addressed by increased acknowledgement of the therapist's participation, influence, human feeling (and so on), is not consequently eliminated: psychoanalytic concepts, whether old or new, serve in part to ease the therapist's discomfort and do not necessarily reflect what he actually does or its effect. In addition, the elevation of the therapist, as I have argued, is to some extent built-in, doubly so: first, the role of analyst is unlike any other and therefore "apart," and second, the psychoanalytic situation works by encouraging a powerful illusion.

THE OFFERING

A peculiar imbalance is built into the therapeutic relationship. By its nature the situation encourages the patient's dependence on and attachment to the (hired) partner, an arrangement that emphasizes the patient's need and vulnerability; the therapist, like priest or doctor or teacher, is in a position of power. Therapy works by offering a sale, delimited illusion, like a play, a paradoxical intersection of the real and the illusory that inheres in the psychoanalytic notion of transference (Freud, 1915; Modell, 1990; Szasz, 1963). The therapist, a living, actual man or woman, is also a presence without explicit dimension, a kind of shadow figured--mysterious, known yet not known, and thus available for the patient's invention. This vehicle of illusion can only be a human being who guides a process of tremendous power that intends its tilt--an imbalance of need and dependence. It is thus a position of both trust and illusion.
    But what does the guardian offer in this peculiar transaction? "It must be said at the outset," Modell writes, "that we do not yet possess a theory that explains how psychoanalysis works" (1990, p. 1)--a startling statement given the endeavor's century-long history, but perhaps not surprising at all given the complexity and ultimate mystery of all human connection. Not only do we not know how psychoanalysis works but what the analyst gives the patient in honoring their contract is also not definitively established and will likely never be; absolute definition is not possible, although we may articulate some components and try to understand the nature of this profound exchange--a transaction, we might say, in human "being." In an essay about the devotional poetry of George Herbert, William Nestrick writes: "The only appropriate gift is discovered to be a thing inseparable from the giver, for the very reason that, in his role as giver, man can only give himself" (1975, p. 200).
    The therapy offering has, like the analyst's role, been variously conceived and named over time. For example, the analyst may offer interpretations, insight, love, respect, reassurance, acceptance, understanding, compassion, or companionship; he may offer passionate companionship, dispassionate companionship, compassionate companionship, all of these, or only some. Indeed, it can be said that both people give and that both receive (although only one pays), and that all of these matters are inextricably bound up with and dependent on each other, just as in any other intimate human relationship. For the immediate discussion, however, it is not relevant whether the one on the serving side of the contract thinks of himself more as participant than reflecting screen, or more as detached than embroiled, nor does it matter what he calls the offering. In every case, the offering is in some sense oneself: with all its imperfection and temporal limitation, it is a mortal gift.
    It is essentially a mystery why any one human being chooses to give the way the therapist gives to another, or why any one cares about (or for) another; even what we mean by "caring" is perhaps in the end elusive. However, the capacities to care and to give derive ultimately from the common condition that separates human beings from animals and gods. The burden -- consciousness and imperfection -- is also the well­spring of the gift. But while this capacity inheres in the humanness that all share, it may only be inherited: that is, one person must pass it to another. To put it another way, the capacity develops only through human connection; it is a gift that must be given and received. The baby's development can only take place "provided one includes with it the care it receives from its mother" (Freud, 1911, p. 220); if the infant is not nurtured, the infant will not live (Spitz, 1965).
    Novick (1982) objects, tartly and rightly, to a colleague's remark that the analyst is the "first person who really understood the patient" (Panel, 1975, p. 175); calling this "the therapist's fantasy of being the perfect and better mother," Novick retorts, "If ... the analyst were the first person to really understand the patient, then the patient would be dead.... No one has ever been or will ever be as important as the primary mother" (pp. 350-351).
    This is perhaps another way to understand Winnicott's aphorism, "There is no such thing as an infant." There is only a baby and (mother) caregiver; without the caregiver, "baby" has no meaning and the baby also dies. With more modern conceptions of reciprocity, we might sensibly reverse the aphorism: "There is no such thing as a mother." The reversal, of course, is not so tidy; although "mother" without "baby" loses its meaning, the mortal woman, unlike the baby, does not necessarily die. And unlike the infant, the patient who loses the caregiver does not necessarily die--although it does happen.
    Psychotherapy is an endeavor conceived and carried out by ordinary beings, its basis the intimacy of human connection. With this observation I return to the commonplace: the therapeutic relationship, specialized though it is, rests on the same foundation as any other close relationship. This sameness of course does not preclude difference; the uniqueness of the psychoanalytic relationship is at the center of the discipline's effort, now over a hundred years long, to say what it is and what it does. However, a problem of great magnitude arises if, in the articulation of what is unique to the relationship, the sameness is denied, obscured, or forgotten. In my view, this fundamental ordinariness is easily forgotten, set to the side as an assumption or psychoanalytic given, in much the same way that Freud buried in a footnote the requirement of the mother's care, an assumption of the provision without which no human development takes place (1911, p. 220, quoted above). The provision is a second human being.
    What of literal mortality? When the therapist dies, what is lost is another soul and one soul will mourn for the lost other. But in order to do this work of mourning, there must be provision for the grieving one and, again, psychotherapy can in this matter be no different from ordinary life. The bereaved patient, however, is in a position of very peculiar isolation, alone, without the person known only in the dual isolation of the consulting room. To quote the poet once again: "How should this grief be properly put into words?" (Horace, from "To Virgil").
    Behind my consideration at every moment of the patient's peculiar difficulty lie two assumptions, the first about a right, the second about a responsibility: When the therapist dies, the patient has the common customary right to take leave of that person; and it is the responsibility of the profession providing the service rendered through the relationship to help with the leave-taking: it is a rite of passage, requisite no less in psychotherapy than in the rest of the world. That expectation is equally part of the contract when a therapist does not die (a therapy ending at its natural conclusion), as if he does die. Although these are matters of common sense, psychoanalysis, whether we look to its theory or to its practical, clinical guidelines, does not reflect acceptance of either the right or the responsibility. "'How reluctant is the professional community to accept responsibility in this essentially clear-cut matter? In 1990, during a Panel held at the annual meeting of the American Psychoanalytic Association, Paul Dewald describes "The longstanding and resounding silence in the field about an issue as commonplace as the analyst's becom­ng ill" (Schwartz, 1993, p. 191). Dewald reports an insignificant return rate when he tried to survey the policies, both formal and informal, of all institutes associated with the Association regarding the illness, death, or retirement of the analyst. Every institute among those who replied declared that the issue was important but that the institute had no preplanned policy, and that the problems could be handled "on an individual basis, given the close-knit functioning of the group" (Schwartz, 1993, p. 192). Both avoidant and self-satisfied, it is a remarkably loose approach.
    These dilemmas of confusion and denial do not arise from an absence of feeling for the patient but rather from a struggle with the power of what the clinician, as well as the patient, feels--a dilemma arising less from an emotional insufficiency than from its opposite, a well of feeling that swamps articulation. Difficulty with the uneasy subject of the therapist's mortality is in part a difficulty of articulation. The technical language of psychoanalysis may be especially unsuited to uncomfortable subjects more readily addressed by poets or philosophers. On the other hand, it may be that psychoanalysis sometimes employs its language in a motivated, if not always conscious, effort to deflect and avoid what might be more fully felt, if it were only more plainly articulated.

THE PROBLEM OF FEELING

And psychoanalysts try mechanically not to be mechanical. They try to show they are human. How can we show we are human? Either we are human or we are not. (Gardner, 1989, p. 4)

 Here I address the problem of feeling--not the patient's but the therapist's. In the first section I have described the silence and confusion that have characterized theoretical discourse about termination. If the heart of the termination process involves the work of mourning, and if therapists are uneasy with their feelings, then we may understand better the silence around termination, both when it is routine and when the mourning is intensified because the therapist is sick and may die. It is my assumption that this mourning and intensification of feeling refer to the experience at parting of both partners in the therapeutic dyad.
    I want to emphasize that always implicit in my consideration of the dramatic example of the therapist's illness or death is the therapist's merely human limitation of every kind; that limitation, however it is manifested, is bound to collide with the patient's fantasied illusion and idealization of the wished-for protector. In such collision, perhaps especially strong in the ending phase, is the heart of the work all along: that is, it effects the gradual dismantling of transference illusion. <>Dis­illusionment (and de-idealization) will thus be a necessary part of every ending as the patient gives up (though never relinquishing completely) the "transferred, early omniscient, omnipotent, all-protective and all­giving parent" (Novick, 1982, p. 35 I ).
    Reviewing the termination literature, Novick summarizes in one paragraph "all there is in the vast literature on termination in regard to the feelings of the analyst when the patient leaves" ( 1982, p. 356). He notes that, at the end of a panel on termination in 1975, one participant remarked on the silence throughout the discussion itself about the subject of the analyst's feelings at the end of a therapy; he summarizes the participant's comments: "the discussion omitted any reference to the sense of loss that analysts experience at the prospect of separation from patients with whom they have had a prolonged, intimate, and frequently very gratifying experience" (p. 355). The single reference Novick presents of an analyst clearly moved by strong emotion at the parting is Balint's admission (1950):

Usually the patient leaves after the last session happy but with tears in his eyes--and, I think I may admit--the analyst is in a very similar mood. (p. 197)

If this virtual silence means merely that analysts' feelings at the parting are inconsequential or nonexistent, then we must conclude that the psychoanalytic relationship involves a prolonged, intimate connection between two human beings, only one of whom is subject to ordinary feeling. As a measure of what actually faces the analyst at the end of the relationship, however, Novick closes with the suggestion that "it may often be ... that termination will be a greater real loss for the analyst than it is for the patient" (p. 356, my italics). That is because, while the patient (in Novick's view) has relinquished mainly the transference object of illusion, a shadow puppet, the analyst knows the patient more fully and clearly, as well as intimately, and therefore loses a more real object.
    It isn't necessary to agree about what kind of loss, real or fantasized (or what proportion of each), is involved for each partner to appreciate the point: that is, the analyst experiences a significant loss at the end of the extended, close relationship with the patient, but analysts have an unusual difficulty acknowledging this sense of loss. That these reactions at ending are rarely acknowledged, much less freely articulated by the doctors of feeling, is a legitimate cause for wonder given that the therapist serves as guide to a patient whose essential task in ending is to acknowledge the loss of the partner. Mourning is at the center of the termination process.
    I will try to illustrate this idea and the difficulty very simply, not by citing the results of scientific studies, but by presenting two examples of therapists who seem uneasy with a range of emotions. Stephen Firestein's book, Termination in Psychoanalysis, frequently quoted in the literature, appeared in 1978. Firestein presents eight case examples which he studies in relation to their termination. Among the questions he considers in a chapter titled, without irony, "Substantive Conclusions" is the following: What is the impact on the analyst of approaching termination? The study informs us that analysts do indeed experience something when a therapy ends, but nowhere in the brief analysis of responses does Firestein use the words "emotion" or "feeling"; rather, he refers to "the impact" of the ending and to "a range of affective responses." He then carefully separates those affective responses into two categories.
    The first category of analysts' responses involve anxiety, and within it Firestein specifically notes reactions of "concern about the stability of the therapeutic result," "doubt about the timing of the termination," and "misgivings about the thoroughness of the work" (pp. 214-215, my italics). While the words naming the categories of anxiety have an affective flavor--"concem," "doubt," and "misgiving"--there is no reference to the relationship; instead, the object of each anxiety concerns the analysts's professional performance and expertise, thoroughly omitting the second person. The second and more problematic category of experience Firestein struggles mightily even to name. The analysts, he writes, "experienced not only varying degrees of anxiety over termination, but gradations of what, for want of a better description, could be called grief" (p. 214, my italics). These words reflect the heart of a difficulty.
    In this sentence Firestein is proposing that, in the year 1978, there does not yet exist in the English language a term adequate to describe the eight analysts' affective responses to the loss of their patients at parting. We must conclude that, truly, they are an extraordinary set, these psychoanalysts whose human feelings defy our common language to describe! A thoroughly reasonable man, Firestein proposes a compromise to avoid a distracting fuss over the limitations of English: He writes that the word "grief," wanting though it may be, will do. The implicit alternative would be to coin a word whose definition might read: "that affective response peculiar to a psychoanalyst at the termination of a treatment."
    Within this category which, until the better word is found, we will also call "grief," Firestein includes "regrets about ending," regret being as close as the analysts venture toward acknowledging a direct sense of loss for a person deeply known. Such regrets surface, Firestein reports, either because the patient may have been an especially "interesting, witty, or gifted individual" (p. 215) or because he or she was "responsible for an especially instructive learning experience for the analyst" (p. 214)--the patient serving chiefly, it seems, as a source either of entertainment or edification. Firestein allows as well the experience of gratification at the conclusion of an analysis, noting that "the analyst's gratifications ... consist of more than the functional satisfaction of doing one's job well." Of what does that "more" consist? Inching closer to personal sadness at losing the person with whom one has worked long and hard, Firestein concludes his discussion: "Separation reactions are experienced by both members of the dyad" (p. 215, my italics). However, here the analysis of responses ends; there is no further consideration of the nature of these "separation reactions," already in the wording somewhat diluted and shifted away from their focus on the analyst by mention of the second "member of the dyad."
    The author's effort here seems Herculean as he works to avoid words like "feeling," "sadness," or "affection" (the avoided range includes, of course, words such as "relief," "joy," "hatred," or anything else common to human experience), and as he in effect apologizes for his (perhaps indecorous) use of the word "grief"; nowhere is presented a therapist comfortable with missing someone grown familiar, just as he or she misses, and easily would say so, other people once present in one's life, from the retired mailman to the former teacher to absent loved ones, whether dead or gone around the block. The analysts appear to need protection from the power of feeling, perhaps less from the patient's feelings directed toward them than from the power of their own. In such extremes, the therapist remains sheltered behind a mask, untouched, the "countertransference" well under control, if not altogether denied, as he safely brings the relationship to its conclusion. "The enthusiastic and engaging assertion of an older colleague many years ago," writes Leo Stone, "that his patient would have developed the same vivid transference love toward him `if he had been a brass monkey,' is alas (or perhaps, fortunately!) just not true" (1961, p. 41). In Firestein's study of termination, the brass monkey is alive and well. "Harold Searles, for interesting contrast, in 1959 refers to the feelings potentially evoked in the analyst at termination: "The real and unavoidable circumstance of the closing analytic work tends powerfully to arouse within the analyst feelings of painfully frustrated love which deserve to be compared with the feelings of ungratifiable love which both child and parent experience in the oedipal phase of the child's development" (1965, p. 300). Accompanying this love, Searles writes, may be "feelings of jealousy, anxiety, frustration-rage, separation-anxiety, and grief" (p. 300).
    In a ground-breaking study, the first to explore systematically the patient's reactions to the therapist's death (Lord, Ritvo and Solnit, 1978), the authors report a broad range of mourning reactions in 27 participants whose analysts had died. Only two of the respondents appear to have experienced no mourning reaction at all following their loss, and one of these, a young candidate in training, is quoted as saying that "the analyst only mattered as a transference figure ... a peg on which to hang attitudes" (p. 193). Although not categorized by the authors as a normal response, this trainee's reported absence of grief is still quite remarkable; presumably, one peg may be replaced with another, and it would indeed be much easier to feel no grief for the loss of a piece of wood than for the loss of a beating heart. The candidate's "peg" has even fewer distinguishing characteristics than Stone's brass monkey: one brass monkey may not so easily be substituted for another, after all, as one peg for another. The candidate's comment makes as much sense as to say that one's parents are handy pegs for one's affection and dependence, the implication being that attachment to the person is neither reasonable nor relevant.
    But my second example offers to the "wooden peg" analyst a worthy companion. Four years after the publication of Firestein's book, Viorst (1982) looks in greater depth at analysts' responses to termination, directly conceiving the separation as a loss in the title of her paper: "Experiences of Loss at the End of Analysis: The Analyst's Response to Termination." Viorst's study is conducted by unstructured personal interview rather than the questionnaire format used by Firestein, and this difference itself accounts partially for the range and depth of her findings in contrast to Firestein's. In her note on procedure, Viorst writes that "the interview setting ... evoked in a number of analysts not only thoughtful comments but a confessional, emotionally charged response" (pp. 400-401). What did they need to confess?
   In marked contrast to the superficial, mild and ultimately vague responses reported by Firestein, Viorst found that "anger and guilt and frustration and disappointment, along with sorrow, were openly manifested as we spoke of loss at the end of an analysis" (p. 401). Viorst believes that the sixteen analysts spoke, for the most part, "with a high degree of frankness and that they spoke individually in ways they would never speak publicly, or collectively" (p. 401, my italics). The italicized words speak to the gap between cultural discoursed--the public and collective face of the discipline manifested in theoretical formulations­-and the actual individual responses of working analysts, responses which they keep hidden.
    Lord, Ritvo and Solnit's study (1978) reports a psychoanalytic candidate who did not need to mourn the loss of an analyst who, said the candidate, functioned merely as a "peg" for transferences; similarly, Viorst reports in her study a spiritual twin of this unbereaved candidate, the only analyst interviewed who "claims no problems with loss at termination. `Professionalism,' he remarks, `defends against loss"' (p. 402, my italics). While the student is protected from feeling by the analyst's role as transferential hook, the analyst in this example is protected by his professionalism. It seems reasonable to view these two, student and analyst, as captive to theoretical formulation turned dogma, and thus imprisoned, cut off from acknowledgement of, and consequently perhaps even access to, the most fundamental of human emotions. But how is it possible for such a therapist to do what the patient requires?
    In his paper "On the `Unexpected' Termination of Psychoanalytic Therapy" (1982), Limentani addresses the effects of sudden endings of various kinds, emphasizing the seriousness of such breaches of expectation. "As psychoanalysis must rank among the most demanding, stressful, and intimate of human relations, where trust plays an essentially central role, we should take it for granted that severe and deep reactions will occur when it is brought to an abrupt end" (p. 420); a little later he writes of concern about "inevitable scarring" and "broken trust and promise" (p. 420). Although Limentani's focus is ostensibly on the patient, these words reasonably may be applied to both members of the dyad, an implication residing in the vagueness of reference. It is noteworthy that, of the seven case examples of forced termination presented, only two actually refer to therapist-induced endings while the remaining five turn the tables, the patient precipitating the ending. Here the unspoken focus is clearly on the therapist as recipient of the unexpected, a fact that reemphasizes the vagueness as to which member of the pair the author means when he speaks of "severe and deep reactions," "inevitable scarring," and "broken trust and promise."
    At the end of the paper Limentani reminds readers of a consideration that he says should always be held in mind by any therapist working with a patient whose therapy must come to an unexpected ending. He is referring to the reactions of people whose therapists leave them, for example, by moving to another city, or by dying. In what is perhaps another instance of implicit emphasis, an omission worth noting, Limentani writes of the special difficulty for the therapist in working with a patient who has been abandoned this way:

Human beings are not reasonable, even though it would be convenient for the therapist if they were. They do not easily accept "rejections" by fellow members of the human species, no matter how unavoidable and totally un­expected the circumstances. (p. 439)

This wry acknowledgement of human irrationality does assign that qual­ity to the patient--that is, to one half of the dyad, while remaining silent about the other.
    Perfectly reasonable patients and, to include the omitted, missing half of the dyad, perfectly reasonable clinicians, would be exactly like the Houyhnhnms Gulliver meets on his travels, creatures who, when friends or relations die, express "neither joy nor relief at their departure" (Swift, 1963, p. 287). Houyhnhnms feel no affection for their young, but care for them according to theoretical principles derived "entirely from the dictates of reason" (p. 286). In that idyllic and chilling Swiftean world, common human feeling does not exist. Only a Houyhnhnm-therapist (and therapists, Limentani reminds us, are "fellow members of the human species") who thus "rejects" the patient by dying perhaps does not discover "the least regret that he is leaving the world" (p. 287). That is, the hypothetical Houyhnhnm-therapist, a perfectly rational creature, not only resembles Firestein's therapist who barely feels the loss of the patient, Viorst's therapist who is shielded from loss by "professionalism," and the candidate whose analyst's death evokes no sense of loss. On top of that, the Houyhnhnm-therapist who does not miss the man or woman who once came to talk to him also does not miss himself, discovering as he is dying not the least regret at his own departure. That is how thoroughly reasonable this creature is. But in a world lacking common mortal feeling, there is no need for the therapeutic exchange--a transaction in human "being."

MASTERING THE LANGUAGE

In his prose book Vigil (1997), the poet Alan Shapiro tells the story of his sister's final four weeks of life, the family gathering during this time to tend her in a hospice room. ]n a section titled "The Doctor," Shapiro tries to make sense of her oncologist's emotional withdrawal from her as she lay dying, of his "fumbling with emotions that squared so awkwardly with his professionalism" (p. 55). Of his sister's deep dependency on the doctor, Shapiro writes:

More than anyone else, early and late in her disease, he determined how she felt about herself. When she was doing well, meaning when she responded well to the treatment he prescribed, he bolstered her self-esteem by saying he was proud of her, she was his best patient, she was his favorite patient. But if he were the benevolent deity when she was doing well, responding to his treatment, he became the deus ahsconditus when the cancer had metastasized, and it was clear she was going to die. (p. 53)

    Later, after his sister's death, Shapiro is faced with the failure of his brother-in-law, also a doctor, to acknowledge the death. At first uncomprehending, Shapiro chooses, in the end, to understand both of these failures not as the doctors' incapacity to feel but as the incapacity to know what to do with how they feel. He writes of his sister's doctor:

Though it seemed he struggled to maintain his superior role as doctor at my sister's bedside-checking the morphine pump, examining her chart--her status as a treatment failure forced him to face her not as doctor but as a needy, fearful, fellow human being, as just another mortal citizen among the dying. She was asking him, as I had asked my brother-in-law, to speak a language that he couldn't master. (pp. 58-59, my italics)

    How are we to help the patient with the task of every termination--to mourn the losses at ending? And if the therapist is dying, how are we to help now with the double task? This patient must mourn not only the end of the treatment and loss of a rightful expectation, that is, the loss of the never-to-be-completed therapy, but he or she must also mourn the impending ultimate loss, inside the therapy frame and in the real world, of a human being. The patient has to say good-bye and will need help.
    The psychoanalytic situation is carefully structured to induce an extraordinary intimacy that is its reason to exist; it is through this singular human connection--an intimacy that intends separation--that the work is accomplished and the endpoint reached. The patient is in a quandary, however, if the helper is shy of endings and also lacks a vocabulary, as Shapiro puts it, to know what to do with how he feels. The capacity to consider one's mortality, by which I mean human frailty and limitation in every sense, perhaps defines the capacity to be a good guardian of the therapeutic situation: a medium through which patient and therapist alike may discover, and discover again, how closely related are the workings of grief and love.
    But the helper who is sick faces an additional quandary, the pain and complexity of which touches the deepest terror. I believe that the therapist, no less than the patient, has the rightful expectation of help from his or her profession. Such matters, while they reflect on the individual therapist, reflect first and most deeply on the theoretical and professional supports that are every therapist's own mortal right to have. Silence for the therapist compounds silence for the patient. In how many ways is the sick therapist abandoned? The patient, we may then say, is doubly abandoned.

REFERENCES
Abend, S. (1986), Countertransference, empathy, and the analytic ideal: The impact of life stresses on analytic capability, Psychoanalytic Quarterly, 55, 563-575.
Balint, M. (1950), On the termination of analysis, International Journal of Psychoanaly­sis, 30, 196-199.
Bergmann, M. (1997), Termination: The Achilles Heel of psychoanalytic technique, Psy­choanalytic Psychology, 14, 163-174.
Blum, H. (1987), Analysis terminable and interminable: A half century retrospective, In­ternational Journal Psychoanalysis, 68, 37-47.
Blum, H. (1989), The concept of termination and the evolution of psychoanalytic thought, Journal of the American Psychoanalysis Association, 37, 275-295.
Calef, V., and Weinshel, E. (1980), The analyst as the conscience of the analysis, Interna­tional Review of Psychoanalysis, 8, 279-290.
Chused, J. (1997), The patient's perception of the analyst's self-disclosure: Commentary on Amy Lichtblau Morrison's paper, Psychoanalytic Dialogues, 7, 243-256.
Clark, R. (1995), The pope's confessor: A metaphor relating to illness in the analyst, Jour­nal of the American Psychoanalytic Association, 43, 137-149.
Ekstein, R. (1965), Working through and termination of analysis, Journal of the American Psychoanalytic Association, 13, 57-78.
Feinsilver, D. (1998). The therapist as a person facing death: The hardest of external reali­ties and therapeutic action, International Journal of Psychoanalysis, 79, 1131-1150.
Ferenczi, S., and Rank, O. (1924), The Development of Psychoanalysis, Nervous & Men­tal Disorders Publishing Co., 1925, New York.
Firestein, S. (1978), Termination in Psychoanalysis, International Universities Press, New York.
Fliess, R. (1942), The metapsychology of the analyst, Psychoanalytic Quarterly, 11, 211­227.
Freedman, A. (1990), Death of the psychoanalyst as a form of termination of psychoanalysis, In H. Schwartz and A. Silver (Eds.), Illness in the Analyst: Implications for the Treat­ment Relationship, International Universities Press, Madison, CT, pp. 299-331.
Freud, S. (1893), Charcot, Standard Edition, vol. 3, pp. 11-23.
Freud, S. (1911), Formulations on the two principles of mental functioning. Standard Edi­tion, vol. 12, pp. 213-226.
Freud, S. (1915), Observations on transference-love, Standard Edition, vol. 12, pp. 157-171.
Freud, S. (1915), Thoughts for the times on war and death, Standard Edition, vol. 14, pp. 273-300.
Freud. S. (1937), Analysis terminable and interminable, Standard Edition, vol. 23, pp. 209­253.
Gardner, M. R. (1989), Sell Inquiry, Analytic Press, Hillsdale, NJ.
Gates H. L. (1997), The naked republic, The New Yorker, August 25 and September 1, pp. 114-123.
Hamilton, E. (1969), Mythology, New American Library, New York.
Hoffman, I. (1991), Discussion: Toward a social-constructivist view of the psychoanalytic situation, psychoanalytic Dialogues, 1, 74-105.
Hoffman, I. (1998), Ritual and .Spontaneity the Psychoanalytic Process, Analytic Press, Hillsdale, N1.
Hoffman I. (2000), At death's door: Therapists and patients as agents, psychoanalytic Dia­logues, 10(6), 823-846.
Horace (1994), To Virgil, In D. Ferry (Trans.), The Odes ofHorace, Farrar, Straus, & Giroux, New York.
Hurn, H. (1971), Toward a paradigm of the terminal phase: The current status of the termi­nal phase, Journal of the American Psychoanalytic Association, 19, 332-348.
Kaplan, D. (1997), Discussion of Martin Bergmann's and Jack Novick's articles, Psycho­analytic Psychology, 14, 175-180.
Loewald, H. (1980), On the therapeutic action of psychoanalysis, papers on Psychoanaly­sis, Yale University Press, New Haven, CT.
Loewald, H. (1988), Termination analyzable and unanalyzable, The Psychoanalytic Study of the Child, 43, 155-166.
Limentani, A. (1982), On the "unexpected" termination of psychoanalytic therapy, Psy­choanalytic Inquiry, 2, 419-440.
Lord, R., Ritvo, S., and Solnit, A. (1978), Patients' reactions to the death of the psycho­analyst, International Journal of Psycho-Analysis, 59, 189-197.
Margulies, A. (1998), Discussion of Ellen Pinsky's dissertation, Mortal Gifts: A View of the Therapeutic Relationship through the Lens of Mortality, Paper presented at The Massachusetts School of Professional Psychology, March 5, 1998.
McLaughlin, 1. (1981), Transference, psychic reality, and countertransference, Psychoana­lytic Quarterly, 50, 639-664.
Mitchell, S. (1997), Influence & Autonomy in Psychoanalysis, Analytic Press, Hillsdale, NJ. Modell, A. (1990), Other Times, Other Realities, Harvard University Press, Cambridge.
Modell, A. (1991), The therapeutic relationship as a paradoxical experience, Psychoanalytic Dialogues, 1, 13-28.
Morrison, A. (1990), Doing psychotherapy while living with a life threatening illness, In H. Schwartz and A. Silver (Eds.), Illness in the Analyst: Implications for the Treatment Relations, International Universities Press, Madison, CT, pp. 227-250.
Morrison, A.(1997), Ten years of doing psychotherapy while living with a life-threatening illness: Self-disclosure and other ramifications. Psychoanalytic Dialogues, 7, 225­241.
Nestrick, W. (1975), George Herhert--the giver and the gift. Ploughshares, 2, 187-205. Novick, l. (1980), Negative therapeutic motivation and negative therapeutic alliance, Psychoanalytic Studv of the Child, 35, 299-310.
Novick, J. (1982), Termination: Themes and issues, Psychoanalytic Inquiry, 2, 329-365. Novick, J. (1988), The timing of termination, International Review of Psychoanalysis, 14, 307-318.
Novick, J. (1997), Termination conceivable and inconceivable, Psychoanalytic Psychology, 14, 145-162.
Ogden, T. (1997), Reverie and Interpretation, Jason Aronson, Northvale, NJ.
Panel (1975), Termination: Problems and techniques, W. S. Robbins, reporter, Journal of the American Psychoanalytic Association, 23, 166-176.
Pedder, J. (1988), Termination reconsidered, International Journal of Psychoanalysis, 69, 495-505.
Pizer, B. (1997), When the analyst is ill: Dimensions of self-disclosure, Psychoanalytic Quarterly, 56, 450-469.
Rangell, L. (1966), An overview of the ending of an analysis, In R. Litman (Ed.), Psycho­analysis in the Americas, International Universities Press, New York, pp. 141-165.
Reich, A. (1950), On the termination of analysis, International Journal of Psychoanaly­sis, 31, 179-205.
Rizzuto, A. M. (1998), Sound and sense: Words in psychoanalysis and the paradox of the suffering person, Canadian Journal of Psychoanalysis, 3, 1-15.
Samuels, L. (1992), When the analyst cannot continue, The San Francisco Jung Institute Library Journal, 10, 27-38.
Schur, M. (1972), Freud: Living and Dying, International Universities Press, New York.
Schwartz, H. (Reporter) (1993), The life cycle of the analyst: Pregnancy, illness, and disability, Journal of the American Psychoanalytic Association, 41, 191-206.
Schwartz, H., and Silver, A. (Eds.), (1990), Illness in the Analyst: Implications for the Treatment Relationship, International Universities Press, Madison, CT.
Searles, H. (1965), Oedipal love in the countertransference, Collected Papers on Schizo­phrenia and Related Subjects, International Universities Press, Madison, CT.
Shapiro, A. (1997), Vigil, University of Chicago Press, Chicago.
Silver, A. L. (1990), Resuming work with a life-threatening illness--and further reflections, In H. Schwartz and A. Silver (Eds.), Illness in the Analyst: Implications for the Treatment Relationship, International Universities Press, Madison, CT, pp. 151-176. Silver, A. L. (2001), Facing mortality while treating patients: A plea for a measure of authenticity, Journal of the American Academy of Psychoanalysis, 29(1), 43-56.
Spitz. R. (1965), Symposium: Transference: The analytic setting and its prototype, Inter­national Journal of Psychoanalysis, 37, 386-388.
Stone, L. (1961), The Psychoanalytic Situation, International Universities Press, Madison, CT.
Swift, J. (1963), Gulliver's Travels, Bobbs Merrill Company, New York.
Szasz, T. (1963), The concept of transference, International Journal of Psychoanalysis, 44,432-443.
Tauber, G. (1989), Re-analysis after the death of the analyst, Modern Psychoanalysis, 14, 171-182.
Viorst, J. (1982), Experience of loss at the end of analysis: The analyst's response to ter­mination, 2, 399-418.
Weinshel, E. (1990), How wide is the widening scope of psychoanalysis and how solid is its structural model?, Journal of the American Psychoanalytic Association, 38, 275­296.