Peter Fonagy - On the Relationship between Technique and Theory

"It is my impression that analytic clinical practice is not logically deducible from currently available theory. There are several reasons for this.

First, analytic technique is known to have originally developed on a trial-and-error basis. Freud (1912b) willingly acknowledged this when he wrote: "The technical rules which I am putting forward have been arrived at from my own experience in the course of many years, after unfortunate results had led me to abandon other methods" (p. ). Freeassociation, for example, is acknowledged by Laplanche and Pontalis (1973) to have been "found" (reached empirically), rather than deduced (p. ). Similarly, Klein's (1927) and Anna Freud's (1926) discovery of play therapy could hardly be considered to have been driven by theory. More recently, Kernberg (1975) made the case for his modified technique with borderline patients by referring to what "clinical experience has repeatedly demonstrated" (p. 91) and the incidental findings of the Menninger Foundation Psychotherapy Research Project (p. 82). Similar acknowledgments to empirical derivation were made by Kohut and Wolf (1978, p. ) and Hartmann (1951, p. ).

But most technical developments are based on ordinary daily experience. For example, Kleinian analysts have learned to emphasize the interpretation of defense and to be a great deal more cautious in how and when they interpret envy or destructiveness. Some British Independent psychoanalysts have determined that fostering regression is not as successful as was once hoped (Rayner 1991). Most British analysts have come to give priority to the interpretation of affect and mental state in the here-and-now relationship (Sandler and Dreher 1996).2

Second, innovative clinical procedures may, of course, be theoretically guided. If this were more frequently the case, we would expect practices to have been logically derivable from theory, at least in some instances. Such claims have commonly been made (e.g., Freud 1904, p. ; Kohut 1971, p. 264). The following specific example will suffice here. Gedo (1979) boldly stated that: "Principles of psychoanalytic practice … [are] based on rational deductions from our most current conception of psychic functioning" (p. 16). In fact, his book made the claim that the unfavorable outcomes of developmental problems can be reversed "only by dealing with those results of all antecedent developmental vicissitudes that later gave rise to maladaptation" (p. 21). What sounds like, and is claimed to be, "a rational deduction" is in fact a hypothesis, emphatically stated to disguise the absence of a logical argument to support it. It is one thing to assume that development follows an epigenetic scheme, but quite another to claim that in therapy, all earlier vicissitudes must be dealt with. There is no evidence for Gedo's claim, even from within the self psychological theoretical camp from which the suggestion emanates (Kohut 1984; Terman 1989). In fact, the differences between Kohut's and Gedo's therapeutic approaches illustrate the absence of a deductive tie between the epigenetic model to which self psychologists subscribe and the technical propositions that are claimed to relate to these. For example, Kohut (1984) explicitly recommended that, under certain circumstances, developmental vicissitudes, such as narcissistic traumata, should be left alone (pp. 42-46).

This example is representative of many widely respected claims for the theoretical grounding of recommended therapeutic techniques or principles. For example, Kernberg (1976) insisted that "an important consequence" of his admittedly inspiring and highly original theoretical formulation concerning the nature of borderline personality disorder is that the therapist's active focus must be on the mechanism of splitting "before any further changes can be achieved with such patients" (p. 46). However, Kernberg failed to demonstrate the claimed deductive relationship. From the same psychoanalytic institute (Columbia), and previously from the same psychiatric hospital (Menninger), Schafer (1983) recommended delaying interpretations altogether for long periods (pp. 165-180). In fact, the most exhaustive exploration of the long-term findings of the Menninger Psychotherapy Project could be argued to have overturned many of the findings of the original Kernberg et al. report (1972).3 the existence of quite contradictory therapeutic alternatives indicates that the theory of borderline phenomena proposed by Kernberg, however persuasive, cannot be connected to any singular approach to therapy through readily discernible deductive steps.

Third, analysts do not understand, nor do they claim to, why or how their treatment works (see, for example, Fairbairn 1958, p. ; Fenichel 1941, p. 111; Kohut 1977, p. 105; Matte Blanco 1975, p. 386; Modell 1976, p. ).4 Is it conceivable that such a state of affairs could arise if practice were logically entailed in theory? Surely, if this were the case, a clear theoretical explanation for curative action would be readily forthcoming. The nature of the therapeutic action of analysis is a recurring theme of psychoanalytic conferences, starting, perhaps, with the Fourteenth International Psychoanalytic Association Congress in Marienbad (Glover et al. 1937), where Glover, Fenichel, Strachey, Nunberg, and Bibring crossed swords. Since that time, there has been a symposium on this topic at about ten-year intervals, alternating between the International and the American Psychoanalytic Association meetings. At each of these meetings, speakers have almost ritualistically asserted that the way analysis works "is not adequately understood" (Fairbairn 1958, p. ), or have indicated "an urgent need for further research by psychoanalysts" (Cooper 1989, p. ). The state of epistemic affairs is well summarized in Matte Blanco's (1975) words: "The fact is that nobody has, so far, succeeded in establishing with great precision what the factors are and how they combine with our understanding to produce the cure" (p. 386). If the practice were logically entailed in theory, we would undoubtedly have a clear—or at least clearer—theoretical explanation for therapeutic action.

Fourth, as has already been suggested, psychoanalytic practice in essence has changed little, if at all, since Freud's (1912a, 1912b, 1913) original descriptions in a few brief papers before the First World War. This state of affairs has been classically acknowledged (Glover 1968, p. 115; Greenson 1967, p. 3). For example, Glover (1968) stated: "For certainly, and despite a multiplicity of articles on the subject of technique… no very radical advances have been made in the therapeutic field" (p. 115). Because, traditionally, analysts have not recorded their clinical work, such assertions are hard to prove. However, extensive supervision based on the reported psychotherapeutic process, which forms the core part of psychoanalytic training, serves to ensure that analysts, at least in the course of training, adhere relatively closely to so-called traditional technique. This is not to say that there have been no stylistic changes in analytic technique, but these have left the fundamentals (freeassociation, interpretation, insight, focus on transference and countertransference) largely unaffected. Over the same century, enormous theoretical advances have taken place, so that it is hardly practical to attempt to provide integrative summaries of analytic theories. The discrepancy in the rates of progress between theory and practice is quite remarkable, and would be hard to understand were it not for the relative independence of these two factors.

Technique, of course, has changed somewhat, and I am not suggesting that current technique is identical to that which Freud evolved, or to that which was generated by key formalizers of psychoanalysis following Freud's death. There is no doubt that change has occurred, but current technique is far more recognizably Freudian than current theory. Suggested technical changes have been relatively minor (e.g., the value of early transferenceinterpretation, or of self-disclosure) and not radical (such as the use of psychodrama in place of freeassociation to reveal unconscious representational systems, the abandonment of the interpretation of unconscious content in favor of psychoeducational strategies, or the use of behavioral or cognitive behavioral adjuncts to therapy). Radical technical innovations are seen as taking the proposer beyond the pale, as if such modes of intervention could no longer be considered to fall within the domain of psychoanalytic theoretical explanations.

But of course, psychic change needs to be explained, whatever its cause (Fonagy 1989). If the current argument is sound, change brought about through the application of classical analytic technique is no easier to account for than change following behavior therapy or religious conversion, and the "inseparable bond" between theory and practice can be maintained only through powerful rhetorical claims. The tendency to disguise the loose coupling of theory to practice behind rhetoric is pernicious because it serves to close the door on imaginative clinical exploration by fostering an illusion of a theory-based technical certainty: "We know what needs to be done because we know how it works and why." Furthermore, the converse is also true. New theoretical ideas can claim acceptance and legitimacy in public theory through a tracing of their origins to relatively unmodified therapeutic technique, thereby reinforcing the immutability of the latter. The slow development of analytic technique is, I believe, in part attributable to the tendency of inventors of new theories to seek validation for their hypotheses via the congruence of new ideas with accepted clinical practices. The practices are claimed as uniquely effective and unchangeable, at least until a new theory evolves.

Fifth, the thorny issue of therapeutic effectiveness might also imply an independence of the domains of theory and practice. There is relatively little evidence to support the clinical claims of psychoanalysis as a viable treatment for psychological disorder (Fonagy, Kachele et al. 2001; Fonagy and Target 1996; Gabbard, Gunderson, and Fonagy, in press; Roth and Fonagy 1996). There is much stronger support for many of its theoretical claims (e.g., Bucci 1997; Fonagy, Steele et al. 1993; Westen 1999), including those related to the treatment process (e.g., Luborsky and Luborsky 1995). While accepting that a lack of evidence for effectiveness does not imply a lack of effectiveness, the discrepancy may also be explained by the assumption that practice is not entailed within theory. The evidence that exists is for a theory of mind that contains unconsciousdynamic elements. Evidence is, however, lacking for the translation rules for moving from psychological theory to clinical practice.

For example, work from other laboratories and mine has provided good evidence for the psychoanalytic notion that a parent's experience of having been parented is transmitted to the next generation (e.g., Fraiberg, Adelson, and Shapiro 1975), determining aspects of the nature of the child's relationship to that caretaker (Fonagy, Steele et al. 1993). There is far less evidence to suggest that addressing the parent's past conflicts in a psychotherapeutic context might help him or her to establish secure attachment relationships with the child(van IJzendoorn, Juffer, and Duyvesteyn 1995). Actually, the theory says little about how knowledge concerning transgenerational relational links may be most effectively used in a clinical context. Does it necessarily follow from analytic theory that insight by the parents into their own childhood experience would be the best way of preventing transgenerational transmission of maladaptive patterns of relating? Or is the closest analogue to insight-oriented psychotherapy chosen by analytic clinicians almost automatically, since this is what serves to define their theoretical identity?

Sixth, as has been implied, it has been impossible to achieve any kind of one-to-one mapping between therapeutic technique and theoretical frameworks. Interestingly, it is as easy to illustrate how the same theory can generate different techniques as how the same technique is justified by different theories. For example, Campbell (1982) demonstrated that clinicians with broadly similar theoretical orientations differed in the extent to which they adopted a position of technical neutrality, shared their thoughts and feelings with patients, or gratified their patients' primitive developmental needs. By contrast, it is equally striking to observe that clinicians using very different theoretical frameworks can arrive at very similar treatment approaches. For example, Kernberg's (1989) work with borderline patients has much in common with the work of those who practice according to a Kleinian frame of reference (Steiner 1993). Both these observations imply that practice is not logically entailed within theory.

Seventh, one may legitimately ask the question: What is psychoanalytic theory about if it is not about psychoanalytic practice? The answer is that it is predominantly about the elaboration of a psychological model, and the way in which that model might be applied to the understanding of mental disorder—and, to a lesser extent, to other aspects of human behavior (e.g., literature, the arts, history, and so on). Freud's corpus may be an eloquent example: his technical papers take up far less than a single one of the twenty-three volumes of his collected psychological writings. The value of theory for the analytic practitioner consists in elaborating the meaning of behavior in mental state terms that can then be communicated to the patient. How such elaboration is done—or indeed, whether it is helpful to do it—is not readily deducible from the theory."　(pp. 19-26)

　

Peter Fonagy (2003). Some Complexities in the Relationship of Psychoanalytic Theory to Technique. Psychoanal. Q., 72:13-47