Gian Paolo Scano

Interaction as a formal object of psychoanalysis

The term "interaction" was kept out of the psychoanalytic vocabulary for an entire century: it was common view that it simply defined "what must not be done", "what is not psychoanalysis." Such a radical ostracism, which is easily understandable from a historical point of view, has a paradoxical flavour: as a matter of fact, any therapeutic action is necessarily produced through an interpersonal relationship, therefore it implies an interaction! This did not slip D. Rapaport's critical mind: while approaching methodological issues in the nineteen-forties, he identified, unexpectedly for those times, the "method of interpersonal relationship" as the specific one for psychoanalysis(1). This consideration, which could have fostered a very different understanding of interaction, failed to solicit any discussion or attention. Moreover, Rapaport himself soon became aware of the difficulties in reconciling this perception with the structure and assumptions of metapsychology(2).
However, during the last decade, the notion of interaction seems to have reacquired citizenship(3) and it now enjoys a full rehabilitation in many authors' opinion. The lines of avowed interactionists are broader and broader. They use the concept of interaction to refer to something that is intrinsic to the very nature of therapeutic transaction. Such position is maintained from different angles. At times it is simply an underlining of the fact that whatever the therapist may or may not do, he cannot but act (Gill, 1994;). At other times the interactive assumptions emerge from the reconsideration of the analyst's neutrality and subjectivity (Renik 1993, 1996) or they appear from a re-reading of transference, countertransference and enactment (Renik, 1999). In yet other instances, the notion of interaction is framed within a more generally "constructivist" (Hoffman 1983, 1998), "intersubjective" (Stolorow, Atwood, 1992) or "relational" vision. (Mitchell, 1988). We owe it to M. M. Gill, though, to have offered the more authoritative, clear and defined proposition: having come to a "constructivist" vision, at the end of a long journey which started from topographic theory, he came to acknowledge interaction as "intrinsic to the procedure" (Gill, 1994).
A traditional position is opposed to that of the avowed interactionists: the former takes the term "interaction" for its narrower definition and refers it to the technical modes of intervention and to kinds of therapeutic action which are understood as being in an intimate opposition to the true essence of psychoanalytical practice (Oremland, 1991). Such judgement rests on a double contraposition (action\interpretation, intrapsychic\interpersonal) which is happily taken as the incontrovertible boundary line between what is psychoanalytical and what is not (Green, 1993).
Between these two extremes lies the heterogeneous middle position of those who keep away from strictly theoretical frameworks and stick to clinical practice, thus mitigating any absolute theoretical and technical traditional positions, with regards, for example, to the notions of neutrality or abstinence (Bordi, 1995), and at the same time avoiding any exceedingly radical catchwords.
The status quaestionis, however, is not clear. The interactionists prove very effective when they give their clinical and logical argumentations, but they appear less vigorous and sure when they try to draw proper theoretical perspectives. It so happens that what risks to prevail in this discordant mixture of voices is the unrelenting work of those who struggle to digest the "intersubjective foreign body" in their effort to reduce it to the already known, according to a common and repetitive way in psychoanalysis.
This work aims at developing M. M. Gill's thesis. Here the notion of interaction is taken in its "stronger"sense, which is alternative to the traditional conceptual frame of reference. It is argued that it should be considered as the actual formal object of psychoanalysis as well as the focal point of the necessary intersubjective reformulation of the theory. As a first step in this direction, a distinction is introduced and conceptualised between interaction and meta-interaction, with the aim of defining two different domains, underlying different classes of variables assumed to be active in the process of change.

1. What is "interaction"?

The term "Interaction" is not a technical one but an everyday one: we need a definition and a conceptual framework if we want to add it to the theoretical lexicon.
In a purely phenomenological sense ("therapeutic" or "analytic interaction"), it is used to address the whole of the transactions that take place between a T (therapist) and a P (patient) and to define what an external observer could describe as taking place during a session or a therapy. Such understanding of the term is acceptable and accepted, but generally deemed as being of scarce relevance.
Instead it is preferred to use the words "interaction" and "interactive" to designate a specific technical mode which is characterised by "active" interventions from the side of the therapist, that is to say, by the adoption of a method which is generally considered alternative to the psychoanalytic one which, instead, should stick rigorously to the intrapsychic, to interpretation, and to neutrality. This second, broadly accepted understanding of the concept has monopolised the debate, causing it to turn into a technical debate focusing mostly on the admissibility of interventions of a so-called "interactive" kind. A technical problem would, of course, also have a theoretical side dealing with the compatibility between traditional theory and interactionist positions which should be harmonised to the customary conceptual frame in order to be accepted.
Although it is perfectly logical from the point of view of those who take the term interaction in this narrow sense, this position is responsible for a misunderstanding: To them it would seem that the interactionists were simply promulgating the rightfulness of what - at different times in the history of psychoanalysis - has been defined as "active", "educational", "suggestive", "directive", "non-controlled" or "acted out". It would therefore seem as if they were justifying forms of intervention that the psychoanalytical tradition has always considered as extraneous to its method. Hence the misunderstanding. The interactionist position is actually a different one: in their view, the notion of interaction does have a technical component, but it mostly carries a specifically theoretical and conceptual meaning. This has been stated in a clear and conclusive way by M. M. Gill: "I regard the interaction in every aspect of the psychotherapeutic situation not as a contaminant but as intrinsic to the procedure. I regard the ideal not merely as unattainable in practice but as fallacious in principle." (Gill, 1994, p. 62)." This assertion highlights the contradiction that lies in the traditional contraposition between interpretation and interaction. In fact, apart from the analyst's intentions, it "...should make clear that it is impossible to make an interpretation that is not at the same time a suggestion or, otherwise put, an action. Since the same is true for the analisand's influence on the analyst, the analytic situation can be described as an interaction , or as Stolorow and his colleagues (1987) put it in the light of constructivism, an intersubjective interaction." (Gill, 1994, p. 52).
Gill is certainly dealing with technical issues, but the implications of his observation go well beyond such narrow domain. If indeed interaction is "intrinsic to the procedure " and if the therapist - apart from his theoretical vision and from the contents of his explicit communication - cannot not interact, then the term must necessarily underscore an unavoidable characteristic of all transactions between subjects, an intrinsic form of human communication and human relation, itself implying the action of a class of variables (pertaining to this level), which cannot be separated or excluded by means of technical or theoretical convenience. It follows that this strong meaning of the term "interaction" can be used to include the totality of transactions (aware, unaware, intentional, non-intentional, explicit, implicit, verbal, non-verbal, emotional, behavioural...) which take place between a therapist and a patient in both a synchronic and a diachronic direction. This interpretation entails a phenomenal-descriptive denotation; it avoids the reduction of interaction to a merely technical fact and aims at emphasising the idea that therapy consists essentially of this interaction, which therefore includes the full range of variables and factors that are responsible for change. Such a conception is a definitive valorisation of D. Rapaport's statement which leads to the identification of the "method of interpersonal relationship" as the specific method of psychoanalysis.
It is reasonable to think that a "strong" interaction can cause changes in the technique and theory of technique, with regard for example to the concept of neutrality, of interpretation or of transference, but, albeit important, this is not its primary focus. Rather, it refers to the essentially interactive nature of human transactions which - being intrinsic to any possible intersubjective relation - cannot but characterise the therapeutic relationship, that is above all an intersubjective relationship.
At this point, two problems arise: first of all it is necessary to understand the reasons for the devaluation (or the complete exclusion) of the interactive factors in traditional theory from a historical and critical point of view. Secondly, if we accept that the problem of interaction is not one that pertains to technique and that it does not deal with procedures per se, but it instead that it refers to a field of variables that pertain to a different logical and factual level, then - from a theoretic-conceptual point of view - it is necessary to start building a model that includes and makes room for verifiable projections.

2. A misunderstanding that is not fortuitous

The stronger sense of interaction finds scarce room within the debate. Something similar happens in a closely related field, that of enactment. Here too, two definitions can be clearly distinguished: a "strong" one - argued for by O. Renik (1997) - which treats enactment, in the singular form, as an unavoidable and ubiquitous characteristic of any transaction, and a "weak" one (Chused, 1991), where enactments are understood as "events" that either can or cannot take place. The discussion on their evaluation, significance and acceptability adopts an essentially technical and theoretical-technical point of view. A singular enactment (which seems intrinsic to the procedure and therefore very close to the strong sense of interaction) is also taken into scarce consideration.
There is nothing accidental in this difficulty to understand a "strong" interaction. It depends on the viscous inertia of the traditional theory of technique, which was codified throughout the course of the 20th century on the basis of classical theory which, due to its epistemological foundations and its conceptual architecture, could not include interactive variables. In fact, it is from that framework that the fundamental methodological assumptions - such as the intrapsychic point of view, the role of interpretation and the assertion on neutrality - are derived. Lacking a clear critical awareness of this logical derivation, the interactionists' issues are automatically weighed against the usual assertions of technique and theory of technique, these obviously being considered as permanently acquired. This is understandable, but it is not logical. If we assert that interaction is "intrinsic to procedure", we are in a fairly obvious way suggesting that the variables which are traditionally indicated as responsible for change can no longer be considered as independent but rather interconnected, perhaps even subordinated to a different class of variables of an interactive order which are unavoidably moved by the sheer exertion of method and technique. This raises a problem which is neither technical nor clinical, but utterly theoretical; one to which we cannot logically reply with technical considerations or by simply resorting to tradition. The traditional technical assertions, in fact, stand on the basis of that very theory, which is put into discussion by the problem that is raised. If the value of an interpretation lies not only in its content but also and mostly in its interactive meaning, it follows that: a) the theoretical assumptions which for a century have guaranteed the predominance of interpretation, fail to consider the action of an entire class of variables; b) the intrapsychic assumption cuts out and arbitrarily reduces the object of study; c) a therapist cannot be neutral, therefore the theoretical grounds that justify such an assumption must be considered contradictory.
The difficulty in approaching this, as well as other great problems, comes directly from the way in which the crisis of metapsychology was managed in the seventies and eighties and, more generally, from a scarce historical-critical analysis which, apart from the great rapaportian season, was never carried out in sufficient depth in psychoanalysis.
When faced with the dreaded vanishing of formal theory, our discipline reacted with a fearful retreat into clinical practice instead of carrying out a close critical analysis of its theoretical, clinical and technical assumptions. Such analysis should have aimed at discriminating what could be used for reformulation from what should have been considered as the purely logical implications of the assumptions of metapsychology. A focus on clinical practice is a commendable thing, but it is risky when it is not matched with a concern for theory. When formal theory was blooming, it could well be left to a few brave ones (Rapaport, Rubinstein, Klein, Gill...) to explore its territory: everyone agreed that the method rested on a theory which was thriving and commonly accepted. After that crisis, on the contrary, the emphasis on clinical practice took on an implicitly defensive meaning, as if shyly and instinctively looking away for fear of facing the emperor's nudity(4). However, this attitude has ended up strengthening the idea that clinical practice and technique are justified independently from the theory that originated them and will keep their value in spite of theory's destiny. A historical-critical analysis can easily show that, for example, the technical assumptions regarding the centrality of interpretation are not based on data, but on the fabric of metapsychology. A similar analysis can show that the distinction between "interactive" and "non-interactive" interventions does not simply control the domain of technique, protecting it from the subjective and intersubjective excesses, but refers instead to the drive theory, to conflict and therefore to metapsychology, perpetuating the delimitation of its point of view, the definition of its formal object, and the subsequent definition of the method.
Psychoanalysis developed as a theory of structure and development of the mind. This theory has logically shaped and predetermined method, defining and justifying - within the frame of its conceptual configuration - the conception of change, model of process, setting of factors and accuracy of procedures. In the Freudian understanding, indeed, P (rather P's mind) is the object of observation, of theory and of intervention. The scientist-observer builds a theory of P which describes and explains his structure, his functioning and his genesis (t\P). This theory makes it possible to determine the causes of P's symptom or of P's life experiences in order to decide "what" to change and "how". This way t\P enables us to determine which procedures and which techniques are adequate for the solution of a given problem and to distinguish correct and effective ones from those that are not correct or effective. Within this framework there isn't any conceptual room for intersubjective or interactive variables, as it is demonstrated clearly by the conceptualisation of transference and of resistance which specifically refer to intersubjective interactions. Since Freud could not rely on any conceptual resources for intersubjectivity he, by means of brilliant ad hoc explanations, reduced one and the other to the intrapsychic, referring resistance to an expression of the same forces that caused the symptom and interpreting P's (transferential) intersubjective life experiences as an autonomous and unrealistic restatement of a past experience (Scano, 1995)(5).
The assumption about knowledge, the intrapsychic point of view and the analytical neutrality originate from this theoretical framework which is markedly positivistic in its premises and structure, where a necessary objectualising caesurae severs the observer from the observed, who is read through the lens of the (physicalist) theory of mind. This theory allows the analyst to go beyond the loud mutability of experience, which, along with symptoms, lapses and dreams, is reduced to an external emergence, to the result and the effect of the factors that are actually active. This way the (subjective and intersubjective) life experience can turn into a "text", defined even in its interpunction, to be decoded through the grammar of unconscious processes and can be related to the actual processes of causation, which, in the end, lead to drive and its vicissitudes. It is the actual existence of such a grammar and of such processes of causation that grounds and justifies a "neutral" intervention and a "true" interpretation. It gives a "scientific" guarantee for procedure, it makes it possible to read resulting data as "clinical proof". All the analyst has to do is to keep his eye untainted through the constant exercise of his analytical neutrality and look at the actual dynamics of the representations which take place in the "intrapsychic". By means of the grammar of unconscious processes, he will decode those dynamics and he will change the "intrapsychic" balance of representations by introducing new connections, so as to change the actual level of causation: the level of energy. The "cognitive" assumption rests on the global frame of the theory of the apparatus, which explains a symptom as being determined by a (ideative and affective) content, which has been removed by a drive conflict. The cornerstone of this conception is the "intrapsychic" point of view that refers back to Freud's strategic move at the time of his shift from the proto-metapsychology of "Project" to the metapsychology of the seventh chapter of his "The Interpretation of Dreams" after he acknowledged the failure of the theory of early sexual trauma with its subsequent heavy accent on the notions of "phantasy".
It follows that the classical Freudian theory is a naturalistic one, based on an epistemology of the "observed object" and seeks to objectify subjectivity as the only way to win psychism over to the realm of science.
From within this conceptual frame, the problem of interaction can only be stated as a technical one. But when we ask if the classical model gives sufficient account of what actually takes place in the course of a therapy and if the factors and processes that it hypothesises are the true active principles of treatment and change, then the problem becomes an urgent theoretical and conceptual one. If the answer is in the affirmative, then the problem of interaction is left as a purely technical one, and even the most anti-interactionist positions of those who - like Green - consider intersubjectivity and interaction as a substantial betrayal of psychoanalytical practice, become perfectly logical (Green, 1993). If, instead, the answer is negative, then the problem of interaction seems to require a Copernican revolution: it would seem to suggest that theory doesn't just need to be reformulated in its more obsolete aspects, rather to be reformulated from the more hidden physicalist and objectualist premises, because the theory would be inadequate to explain therapeutic interaction and actual factors and processes of change. Logically, in this second situation it would be totally irrelevant to refer to usual conceptualisations and to consolidated tradition.
From this point of view the problem of interaction is neither a technical nor a merely theoretical one (i.e. compatible or incompatible with traditional theory), rather it is a theoretical-conceptual one and it requires that we overcome metapsychology, including its hidden and secret traits, through a strategic redefinition of the "shape" of theory. We must also reformulate the issue of the object of theorising and of the factors and processes of change.

3. Interaction as a formal object of psychoanalysis

Taking interaction as intrinsic to procedure (and consequently value Rapaport's statement that the intersubjective relation is the specific method of psychoanalysis), it is possible to look without nostalgia to the old theory: it managed to accomplish its task extremely well. Freud, having grown up in the spirit of the Berlin school, couldn't build anything but a physicalist theory of mind as the formal theory of psychoanalysis(6). From a historical point of view, this was unavoidable and it was heuristically very fruitful: in fact it made the foundation of a realm of knowledge possible. Lacking any proper "theories of mind", Freud turned to nineteenth century psychology more than it is generally acknowledged for theoretical solutions. This allowed him to build a model of the apparatus that was able to explain behaviour, as well as an intermediate psychodynamic and clinical theory, which made it possible to approach individual cases. This way psychoanalysis ended up being a sort of theoretical "singularity", having a pivotal role in the foundation of the field of clinical psychology, producing a first modern theory of mind, a language and a preliminary organisation of knowledge.
Nowadays, however, it is no longer clear neither that psychoanalysis should have a theory of mind of its own, nor that this would be the right form of theory for psychotherapy. Neuroscience produced new and more reliable theories of mind(7) and our knowledge about nature and about the functioning of mental processes is wider. Moreover "baby observation" offered information on infant behaviour that are more reliable and controllable than any conjectures based on analytical data. First and second cybernetic gave rise to great innovations concerning the organisation, interaction and modification of systems (Atlan, 1972, 1979; von Foerster, 1981). Genetic epistemology offers radical new ways of looking at cognitive processes and at the very way of understanding knowledge, invariance and change ( Maturana, Varela, 1980, 1985). Lastly, psychotherapy research produced data that do not seem to confirm traditional hypotheses. For all these reasons, it seems that psychoanalysis can finally take on its specific role as a method of intervention, which assumes its theoretical foundations from those disciplines that "professionally" deal with those fields of knowledge that are useful or necessary to the exercise and the improvement of its practices. This way psychoanalysis could devote all of its energy to its specific theory: the theory of treatment and change through an intersubjective relationship.
The historical-critical analysis, apparently so strict in underlining the radical Freudian reductionism, can all the while show just as clearly that the intersubjectivist vocation rightfully belongs to the oldest heritage of psychoanalysis. From its very beginning there is indeed a bitter (albeit heuristically fruitful) antinomy between the assumptions of theory and the implications of method. While theory was based on a naturalistic assumption of the observed object, both method and clinical practise stemmed from an intersubjective setting where an observer observes the observed who observes. Freud did not build this device anew, he took it from the habitual clinical method as a method for cure, enlarging it and transforming it in the light of his theoretical framework, making it a sort of metaphorical double of the scientific laboratory. So he built an "epistemic device" (E. Funari), which was "clinical" because of its being "epistemic". Unexpectedly for Freud and his theory, however, such a device revealed an underlying, unavoidable characteristic, a meta-rule of functioning that turned it an "intersubjective device for attachment." It is not therefore astonishing that, coherently with its own frame, theory states that change is the effect of interpretation and insight and also that, simultaneously, the practice of method has come to impose an epistemology of the system-patient and of the system analyst-patient which causes a controlling and norming epistemology to short-circuit. This actually imposes, in clinical practice and progressively also in theoretical statements, the consideration of different level factors that connect change to "experience", "relation" and "attachment"(8), although this often happens in an unclear and indefinite fashion. Interactive and attachment factors emerge, indeed, in transference and in those concepts that give room to interaction, although this is done within the formal limits of the intrapsychic language (identification with the analyst's super-ego, projective identification and counter-identification, enactment...). Still, it is hard to deny from a historical and critical point of view, that psychoanalysis and Freud himself always accepted two classes of active factors: the cognitive ones and the ones pertaining to the relationship(9). Indeed it could be argued that the history of psychoanalysis is a perfect example of spontaneous "experiment" in experimental epistemology, with the analytical couple as its subject-object. The analytical couple, through its actual activity, has developed its own inner point of view, its autonomy, and its actual way of "knowing" and "changing", imposing its intersubjective reality even on a naturalistic theory, which would necessarily deny it. This need to break free from the traditional theory could be seen as the result of the kind of development and acquisition of knowledge which took place in the original device. It forced the naturalistic cage to make room for the progressive explosion of intersubjectivity in the unstoppable expansion of transference and countertransference to the entire technical-clinical environment.
To value interaction and conceptualise intersubjectivity is not to force psychoanalysis into new foreign clothes. It is instead to develop an aspect that, existing from the start, has remained unexpressed due to the conceptual and epistemological conditionings of its imprinting. It is renown that the purpose of a good theory is that of dying in the generation of a new one: perhaps the theoretical convulsions of these last thirty years are the symptoms of this necessary, desiderable birth. The therapeutic dyad is not made of an observer and an observed, rather of the interaction between two subjects, where "an observer observes> an observed who observes> where each is at once "observer" and "observed object" with regard to himself. It follows that theory should focus on this essential datum and produce adequate models and conjectures. In this situation, the object of theorising would not be the "mind"(10), instead it would be first and foremost the intersubjective interaction. This would lead to: 1) the traditional intrapsychic point of view should make room for a more logical intersubjective-intrasubjective point of view; 2) the traditional t\P scheme should turn into a more congruent t\P-T; 3) the linear logic of classical theory should make room for a circular logic, where 4) next to an external, objectivating and normative epistemology, there should be room for an "internal" epistemology pertaining to the system and to the subjectual subsystems. In order to give an account of such complex reality, it is necessary to take an "external" point of view, external to the dyad and to both subjective systems (point of view of the observer) and at the same time also an internal one, internal both to the dyad and to the two subjectual subsystems (point of view of the system and its epistemology.) Therapist and patient are like subjectual subsystems of a dual intersubjective system, which must be taken in its complexity, where change doesn't take the shape of complicated problem, rather it is an exemplar case of "complex"(11) problem.
Gill's argument - pinpointing an essential intersubjective and "interactive" twist as being intrinsic to technical practice - becomes the starting point not just for a mere reformulation of some chapters of technique and clinical theory, but for a global reconsideration of the theory of treatment and the starting point for the formulation of new, different conjectures about the factors and processes of change. Here therapeutic interaction, far from being a simple phenomenological-descriptive notion or a technical category, is the very essence of method as well as the instrument and "locus" for its exercise. It follows that interaction poses a problem that is not simply "technical" or "theoretical", rather it is theoretical-conceptual in relation to the definition of the formal object of theory. The therapeutical transaction, the conjectural process of change and the factors active there should, therefore, rise to the level of formal object of theory. Within this view, "interaction" is not a theoretical "object" that must somehow be made to match the theoretical-clinical apparatus: it is simply the "thing" through which change is assumed to take place and which must be explained by means of theoretical conjectures. In other words, interaction becomes a formal object of psychoanalysis.

4. An alternative point of view

This need to modify the general vision of method and theory does not stem exclusively from logical and theoretical reasons. Luckily, for the past twenty years it has been possible to think about psychotherapy in a way that is relatively independent from traditional theories. Indeed, research has produced a certain amount of data which are to some extent embarrassing, not only for psychoanalytic theory and its internal debate, but also for competing theories. These data are crucial to the issue of interaction, specifically: a) the fact that psychotherapies do work; b) the so-called "Dodo verdict"; c) the prevalence of factors linked to the relation and the image of the therapist over technical factors; d) a scarce correlation between knowledge\insight and change, e) a scarce coherence between technical dictates and what actually takes place during a session. Furthermore, these data would seem to correspond to the results of the Topeka research. As much as they are not definitive, they seem quite coherent on the one hand in their not supporting traditional hypotheses, on the other in encouraging a shift of the focus of research towards the intersubjective field.
So far psychoanalysis has not taken these embarrassing data into much consideration. Apart from a minority engaged in research, it has mostly tried to limit the damage by means of the renown distinction between "specific" and "aspecific" factors within the frame of a general greater evaluation of the implications of the relationship. A distinction between "specific" and "aspecific" factors is perfectly logical and understandable within a meta-analysis of the results of research, which show that therapies based on different principles are equally effective, so that "everyone has won and all must have prizes". In this sense, the "specific factors" are those that each therapy indicates as active in its method, the "aspecific" ones are those that the meta-analyst must presume to be active in all different therapies, in order to explain the "Dodo verdict". On the contrary, it is incongruent to distinguish between specific and aspecific factors within a single method, even more so if there is a suspicion that the former do not produce the desired effects and the latter are much more active than theories would maintain. In this case, aspecific factors should be granted a name and be considered "specific". Logically, theories should be reformulated. If not, a distinction would result as an ad hoc expedient to protect old positions, and simultaneously find no matter what explanation for unexpected data that research should bring.
Instead, the growing value given to relationship is probably a given fact, as far as clinical practice is concerned. Even from this point of view, it is not a matter of pushing towards a generic revaluation of so-called "relational" factors. Rather, it is a matter of acknowledging several things together: firstly the logical and theoretical issues, which show interaction to be intrinsic to the procedure, then the important data offered by research, and, finally, the results of the history of psychoanalysis, especially of the ever-growing weight of intersubjective factors that determined an overgrowth of the concepts of transference and countertransference. All this leads to a necessity to reformulate the theory of treatment so as to focus attention on interaction and to take the intersubjectivity of the method as its starting point.
To this end, starting from interaction as the formal object of psychoanalysis, we can draw a new and global approach and consider therapy, the series of sessions or a single session as a sequence of transactions, which, at a semantic level, produce a flow of scenes and ultimately, "a story". This flow of interactions is the phenomenological "thing", which, understood as the means of a change and of a process of change, is precisely what must be explained. At times, mostly in research, this assumption could be turned into an actual fact. In those cases we will not be dealing with the elusive material typical of the discussion on "cases", but with an actual tape, perhaps turned into a written text, where the sequence of scenes is set in a way that is very close to a collection of data, as Dennett, in a similar but different situation, proved to be possible (Dennett, 1993). The tape and its transcription document therapeutic interaction. In a normal therapy setting, such a text will not be produced, and the therapist's position would certainly be different and more complex. Still, one of the tasks of the therapist is similar to that of he who studies a tape, even though he cannot assume any objectivity, even less so regarding a "tape" which is purely virtual. If we analyse such a tape, we will find that it can be looked at from at least three different points of view:

a. Starting from a normative epistemology and an external point of view, the tape can be approached as the "object observed" by an independent scientist-observer, turned into a written text, where the sequence of scenes is fixed in a similar way to a collection of data. This is the level of the observation and of the scientific study of the event, the place where hypotheses and conjectures are formulated about what takes place inside the consultation room. In this light the "interaction" is precisely the "thing" that must be studied and explained. Incidentally, the adoption of this external, eteronomous point of view by an observer-scientist, would finally settle the methodological anomaly of psychoanalysis: the session has always simultaneously been seen as the place for cure, for observation, for research, and for theorising.
b. A second point of view is autonomous and internal to the system, and it looks at the actual interaction between T and P in their being immersed in an intersubjective situation, because of which they cannot not be there, not behave, not communicate, not interact. It is interaction as a fact rather than as an experience or an active choice. As Gill acutely argued, the whole of T's activity - including his interpretative activity- takes, first of all and beyond T's own intentions, the value of intersubjective interaction, laden with meanings pertaining to that level. Here interaction is intrinsic to procedure as a consequence of the fact that the psychotherapeutic method is the method of intersubjective relation. By calling this level "interactive" or "interaction", we wish to underline that the therapist's participation, his being unavoidably interactive, does not depend on his will but - because it is implicit to the intersubjective structure of method and to the nature of the instrument - it is a mere consequence of the intersubjective circle. He can be hypercontrolled or think he is acting as a fundamentally blank mirror or instead as an hyperinterfering directive agent, but he can't avoid "being there."
c. Another point of view is that of the therapist within the flow of therapy: there he takes a "meta" position and makes tactical and strategical decisions that are necessary for him to lead the psychotherapy. He does so by watching the action and reframing it on the basis of his theoretical, clinical and technical knowledge and by means of the models produced by the scientist-observer. We can refer to this role and this activity of the therapist by the term "meta-interaction" (or "meta-interactive" role, position or level). This is to underline that the therapist must necessarily take up a point of view which for many aspects implies his placement in a meta position, an observation, analysis, (logical) judgement and therefore also a direction with regards to the level of interaction. In a way, as far as meta-interaction is concerned, the therapist serves as an internal supervisor in force of his theoretical and technical competence and of his constant monitoring over "what is actually taking place in the interaction", as well as on himself as part of the interaction. It must be added that, within this "meta" dislocation - even though he takes an "external" point of view (from the dyad, from P and from himself), which is somehow similar to that of the scientist-observer - T is still "part" of it so, in spite of himself, he functions according to the circular logic of communication between subjects and not according the usual and linear logic of an external point of view.
d. In fact, the interconnection between these three positions is governed by circularity: the findings of the scientist-observer will be used by T, who will thus be better able to carry out his meta-interactive activity, which in turn will never break free from the interactive position intrinsic to the procedure. As this process unfolds, the scientist-observer will gradually have the opportunity to ask questions that are more and more precise and to obtain new data, allowing him to sharpen his hypothesis concerning the two different classes of variables (interactive and meta-interactive). Thus he will be able to present the therapist with models that are increasingly specific, and yet they will not be applicable other than within the interaction intrinsic to the procedure, and so on. This circularity is threefold: it takes place between the necessary eteronomy of research and the unavoidable autonomy of the intersubjective system (that is to say between research and clinical practise); between interaction and meta-interaction (and between classes of corresponding variables) within the session and in the dual intersubjective system. Lastly, there is circularity also between autonomy and eteronomy, interaction and meta-interaction, inside each of the two subjects that take part in the transaction, who are simultaneously their own "observer" and "observed object".
e. Indeed, in the intersubjective circle, P is also involved in an "interaction": he also, albeit less evidently, carries out a constant "meta-interaction", which is an essential part of the process of his problem solving. This happens through the explicit, silent, or unaware analysis of schemes and through their assimilation or non-assimilation within his interior world. Such meta-interactive activity could, perhaps, be seen as a fundamental element for recodification and it could be taken as the equivalent of those processes that are traditionally labelled as elaboration.
f. Of course interactive elements can be perceived, verbalised or expressed. This can be through experiencing, the telling or the interpretation of life-experience both conscious or unconscious, or through the action of producing a context, theory, inference, meaning, or aim explicit. Such an element necessarily enters the meta-interactive domain. It is likely that this happens only partially and that the more relevant portions of life experience remain, by definition, within the interactive domain, and cannot be perceived or verbalised other than in an indirect way.

The distinction between "interaction" and "meta-interaction" in the intersubjective scenario of a therapy, as much as it hadn't been explicitly introduced so far(12), simply seems to make explicit the general meaning of M. Gill's position, which became increasingly clear in his later works and reached a final formulation in what can be seen as his testament. Fundamentally, it aims at going beyond the term "relation", which is not sufficiently perspicuous, in order to explore the possibility of approaching the intersubjective implication of method in a broader and more flexible way, and to mark the technical territory, to open a conceptual space where those variables can be defined that we generally define as "aspecific". In short, the meta-interactive position looks at the therapist's action from the point of view of his considerations and technical choices and includes the entire field of "technique". Instead, the interactive position refers to the therapist's "being there" as a subject that is involved in a deep relation with another subject and it also refers to his actions and replies as a subject that is fully included in the intersubjective circle, whose actions and meanings are organised around the rules of any similar "structural coupling"(13).

5. Interaction

In the past the therapist's interaction was not considered as a central element of method, rather it was exorcised as a dangerous spurious agent or as an undesirable guest. And still, the matter is not whether it is legitimate, right or technically proficuous for the analyst to take more action, to be more involved or to take a more "active" participation. The problem regards the impossibility for the therapist to avoid an interactive position in spite of his theoretical and technical positions. Interaction is not, in fact, a gradient that T can modulate, increase or decrease according to a situation or to the patient's psychopathological state. What can vary according to these and other parameters is T's "activity" or choice of procedures. The interactive position is, on the contrary, an unavoidable basic ecological niche where any action, communication and behaviour from T and P take shape. Within this niche, T can interact in a way that can be right, wrong, good or bad for the achievement of the desired result. In any case, he will not be able to not interact. It is therefore essential to focus research on interaction in order to identify its "rules".
The rules of therapeutic interaction are not those of the setting, rather they are those of a dual system constructed through the transactions between two "Egos". A setting, (rather, the set of rules that structure a setting), is a control system, a set of boundaries that cannot be crossed. In the game of chess, for example, the rules set the limits within which all possible games will have to be played; they define the procedure that each game will have to follow in order to be qualified as a game of chess. In the same way, each therapeutic match will have to take place within the rules of its setting in order to be considered therapeutic. Just as it is for the game of chess, however, these rules do not determine strategy, tactic or the evolution of the match. These depend - for to chess - on the logic of positions, of attack and defence, in therapy they depend on different levels of negotiation, communication and signification.
Within the limits defined by the setting, the rules of interaction seem to be the same as those of any structural coupling between two "Egos". They should presumably relate to those pivotal points of experience, which we can temporarily call by the terms "embrace", "understand", "accept", "reject", "push away", "attack", "approach" "give", "deny", "return", "state", "confirm", "disconfirm", etc. These pivotal points, as such and relatively independently from their content, are basic channels that have a meaning and a value in themselves and directly solicit reactions and counter-reactions. They can be seen as semantic joints connected with emotional, scenic and narrative characterisations which mark the relation, although they do so from the bottleneck of subjectual codifications, which are built for each subject within his past history of his structural couplings. The actual intersubjective interaction flows through these joints and channel, where - depending on the recurring activation of this or that channel (for example the channels "accept", or "deny" or "refuse"...), on the condivision or non-condivision of the actual notion of the channel in use, on the acceptance or non-acceptance of the mutual marking of the channel (your saying that "you are accepting" is\is not "accepting"), on the agreement or disagreement on its practicability - a specific pattern of the negotiation and a specific configuration of the interaction are drawn in the intersubjective storyline. This way, apart from the "meta" contents and definitions (put forth, for example, through an interpretation), silent "scenes" are interwoven. Both actors continuously mark their significance and values by activating adequate emotions, confirming or disconfirming their expectations regarding the expected responses and offering reasonable patterns of reply. The proper and silent exercise of these rules and the actual negotiation, developing on the chessboard that they outline, modulates the elementary semantic of relations. This flows through channels and joints that are emotionally charged and can provisionally be defined as: to be acknowledged\ignored, accepted\rejected, understood\non-understood, loved\hated, rewarded\punished, approached\push away, dominated\non-dominated, attacked\non-attacked, included\non-included...
Scenes are marked relatively independently from content and from the official and verbal (meta-interactive) definition of interaction. It takes shape out of sight along with the development of any content; indeed it pertains to "what is truly happening now " above and beyond "what you and I say is happening." This marking cannot be reached from meta-interaction. It can be inferred and understood only when it has already happened, a posteriori, therefore it cannot be intentionally modified: at the most it can be contradicted by a new, subsequent, different marking.
Within this frame, "channels" and "scenes" would not be contexts nor contextualisations, instead they could be seen as elementary bricks in the building of contexts, as turning points between attributions of meaning, switches for emotions and feelings and triggers of silent emotional, or even motor, responses. These bricks, by virtue of their constant employment, can come, at a different level, to cement an invariance, to set fixed rules of contextualisation, thus shaping experiences, perceptions, anticipations and, eventually, even "theories" about a given contextual situation or even more general theories about one's own and other's minds. They would also subsequently structure fixed schemes for reactions, for the selection of goals, and the strategies and tactics to reach them(14).
It is reasonable to think that the elementary semantic of channels and intersubjective joints is relatively fixed and shared among all members of the species. It could include a basic common definition of the values of "yes" and "no", the patterns of acceptance and refusal, the pleasant and unpleasant ones; and this on the basis of a connection with primary emotions and more generally as a consequence of the fact of having brains and bodies that are similar in structure and functioning. Even the more redundant and recurring rules that produce "configurations" and "scenes" could be relatively common and based, for example, on a sort of elementary statistic. Finally, the fact of belonging to the same noosphere, to the same cultural environment (and sometimes to the same "cultural dialect"), can also organise such patterns in a relatively homogeneous way.
We could therefore think that an effective psychotherapy is one that favours "positive" joints (to accept, understand, love...) over "negative" ones (reject, not-acknowledge, hate...) and that a good therapist is one who, by contract, favours positive behaviours according to criteria which were sometimes codified ("accepting mother"). The point is that the marking of the scenes is not about what seems to happen or is defined to happen, but it is about what actually happens in the subjectual organisation of each participant, due to their operational closure and the actual negotiation that takes place within the context of the rules of the dual system. The two people who meet in the consultation room have their own structured way of marking interaction and their own given set of rules. This makes it impossible to predefine the criteria for "positivity" or "negativity". In spite of the similarities and isomorphisms indicated above - of which it is difficult to doubt - it seems that contextualisations, expectations or pathological expectation systems (perhaps even non-pathological ones), tend to act as a sort of bottleneck or a one-way traffic system, predetermining the direction and path of the interaction. At times they even get to the point of favouring what -from an external point of view stemming from criteria based on efficiency and logic - could be understood as "negative" rather than "positive" for the subject(15).
We will leave it to empirical research to find the more general prototypical configurations and the most frequent transformations that can change the valence of channels and joints in force of life experiences and of the individual subjectual history, all the while determining recurrent kind of structural coupling. Consequently if we start from reasonable conjectures, it is up to empirical classification to check and define the specific occurrences that characterise therapeutic negotiation, and to find those that, in the various types and classes of configurations, favour a positive or negative development of the process(16). We can easily infer that all of those occurrences will be positive which are a positive perturbation to P's anticipations so as to favour the reopening of negotiation spaces. Perhaps even more so, those that P marks as indicating greater consonance between what is "said" and what is "experienced" in the double direction from P to T and viceversa (what I say is what is, what you say is what is, what we say is taking place coincides with what I "see" as taking place), apart from their positive or negative, pleasant or unpleasant content.
If this understanding of the rules of interaction and of the construction of interactional meaning were to be taken as plausible, then we could come to the definition of an unusual conjecture: the basic rules of psychotherapy, namely the ones that are more directly concerned with the attainment of its goals, would not be those that traditionally build up the basic chapters of technique. Instead they would be the more general ones of human interaction, precisely those concerning the intersubjective web where, in the end, "Egos" are born, develop and supposedly change. This could lead us to overthrow the traditional view on the so-called specific and aspecific factors, and to reach an unforeseen switch of roles: factors that were always considered aspecific would truly be the actual specific ones(17).

6. Meta-interaction

Although it takes place within a given setting, interaction associates psychotherapy to the other human relations, making it a member of the same class. It is not peculiar to psychotherapy: what distinguishes the latter from the other human relations is, of course, the systematic and controlled use of a meta-interactive level, within the binding rules of the setting. Such a level, where T's " meta-interactive position" takes place, pertains specifically to the reasons that caused the therapeutic interaction and to the goals which it intends to reach (P's problem and its solution); it rests on the knowledge that is necessary for a correct definition of the problem (T's theoretical and theoretical-clinical competences); it translates in the adequate technical strategies to reach a solution and in the control over the development of the process. In the end, meta-interaction coincides with the traditional realm of technique, which implies a shift to a level that is "meta" with regards to interaction, an observation, a control and a direction from the specific point of view of the problem, of its solution and of what it is necessary to do in order to reach it.
This observation and constant examination of the "interaction" from the point of view we have come to define, is exercised: a) over the interactive process ("what is actually happening between T and P at a given t moment with relation to the current context and to a diachronic context "); b) over P's manifest and/or inferred experiences and his/her contextualisation; c) on one's "presence", regarding the command of P's context and the actual configuration of the process. Taking the field of the theoretical, clinical and technical knowledge and the field of the setting as our reference and criterion, we can carry out such an examination by constantly "reframing" what is happening in the light of hypotheses concerning P's actual operational and relational modes (according to any formulation in terms of transference, central relational conflict, unconscious level etc.) and of the strategies and tactics of the project of intervention. This evaluation is the field on which T can take, and indeed takes, his decisions about what to do here and now, about what is happening and about what it is useful to do towards the intermediate and final goals of the therapy. That is to say that the decision is about T's action and the actual tuning of his technical choices about saying-not saying/doing-not doing. This is where the various traditional chapters of technique regarding therapist's action and specifically regarding interpretation (of dreams, symptoms, life experience, defence, "transference"...) join method.
In the traditional description of procedures and their correct application, great importance is given not only to these specific aspects of technique, but also and mostly to a series of more general criteria which are considered as intrinsically characterising the correct analytical action, so as to be taken as essential methodological characteristics. In the common understanding, these criteria are: neutrality, abstinence and non-directivity. These carry a value and meaning as general points of reference for a meta-interactive position, as long as they are redefined within the new context. This is not just about avoiding the extremes that such "norms" have often embodied, but it is about placing these criteria within the intersubjective context. In fact, a radical interpretation of neutrality and abstinence can rightfully be taken as a last attempt to take the therapist out of the intersubjective circle, or as a sort of almighty denial of the intersubjectivity of the therapeutic situation and of the unavoidable interactive presence of the therapist.
In its more extreme form, neutrality is, as Gill pointed out, the assumption that the analyst is like a screen or a mirror that only reflects onto the patient what came from the patient (Gill, 1994). This wouldn't just seem extreme, but it would seem an impossible and paradoxical task. Really, such a "neutral" therapist would be a therapist who unwittingly takes up a good part of the "blank screen." In fact he inevitably influences P with his "interactive" or with his "meta-interactive" action, just as he is unavoidably influenced by P. In a more balanced view that accepts the implications of the intersubjective circle, neutrality would seem to be identified with, on the one hand non-directivity (as a tendency to avoid direct or indirect manipulation, as a willingness to analyse any voluntary or involuntary influence over P, and as a form of respect towards P and of his autonomy and operational closure); on the other hand with the possibility for T to take a "meta" position and to grasp what is going on in the therapeutic scene as it is continuously determined by both actors.
Although it is linked with neutrality, the issue of abstinence is more obscure. It refers to Freud's notorious observation that, as far as possible, the analytical cure must be carried out in a state of deprivation, of abstinence. This is meant as a strategic-tactical choice, in fact: "It is suitable to refuse the patient those very satisfactions that he most strongly desires and asks for with the greatest insistence" ( Freud, 1919). The exacerbated emphasizing of abstinence is not only a consequence of neutrality and an implication of the structure of the Freudian epistemic device but it must also be considered as closely related to the drive foundations of his theory, and to the subsequent conceptualisation of action as acting. Actually, since abstinence is a strategic or tactical choice from the therapist's side, it appears as an action, and it is not, therefore, neutral. In a more balanced view, we could perhaps consider abstinence as overlapping with neutrality and with non-directivity, with the addiction of an irrepressible behaviour and a total respect of the setting and its restrictions.

7. Interaction and meta-interaction: method, technique, research

We can now discuss the issue of meta-interactive elements and their relation to interaction for the purpose of giving a general definition of the relationship between the two domains before we produce a preliminary evaluation of the advantages that such an approach would offer.
The characteristics that are peculiar to "interaction" and to "meta-interaction" allow a simple description of their more relevant articulation. The fundamental quality of interaction is that it "happens" and it cannot be erased or modified by meta-interaction (it can betray it or falsify it but it can not make it unhappen). Simultaneously, though, interaction cannot be grasped or reported (to oneself or to somebody else) other than through a meta-interactive operation. The essential element of meta-interaction is that it always implies an interaction in the sense that, as we said, any interpretation, apart from its content, happens within a context as an action with meanings of its own, which are not necessarily those intended by the agent. As a consequence, within his technical or meta-interactive activity, T inevitably interacts. This can confirm or disconfirm the points of junction between the marking of the scenes and P's basic contextualisations by providing him with further proofs of his "theories" at the very moment when he is busy dismantling them. Conversely, it can introduce useful perturbations which may enable him to modify them.
This approach seems to suggest a general conjecture concerning the factors and active principles of change: the "meta-interactive procedures " - that is to say the entirety of the traditional technical aspects (including interpretation, transference analysis, resistance analysis etc.) and their immediate effect in terms, for example, of "self-knowledge" or insight - should not be considered as the actual factors of change or as directly connected to factors of change. Instead, they should be seen as favouring an effective "interaction", which is the place where we should look for the actual factors of change. In this sense, these procedures are what makes the interaction between T and P able to work properly as a therapeutic interaction. This conjecture relies on the assumption that T cannot, by means of linear and direct causality, modify P, who can only change himself from inside, starting from the possibilities and limits posed by the structure of his systemic organisation. T can only favour the process of change by stirring the relational and semantic level through an interaction which offers different answers from the ones that P's semantic organisation would expect. He will not be able to take up this role by virtue of special and uncontrollable internal attributes, such as for example an extraordinary relational competence or an extraordinary ability for "unconscious communication". He will be able to do it through a correct, rigorous and constantly reiterated meta-interactive activity. Of course, in a few lucky situations, T could have such extraordinary abilities, but such a case would be statistically irrelevant and practically unforeseeable. A therapist with an average relational competence and adequate training, could carry out the task (albeit not in every possible case), through an appropriate, careful and continuously tested meta-interactive activity.
From this point of view, the task of the attainment of the desired change depends essentially on interactive factors, and more precisely here, on P's reaction to the perturbation of his semantic environment and of his usual emotional codifications. Instead, the rigorous and proper exercise of meta-interactions takes the substantially instrumental - but essential and irreplaceable - role of making an effective positive perturbation possible and more likely in the various situations of the process. Such a control system ensures that an effective interaction is not merely an accidental, statistically irrelevant event.
The meta-interactive position has a further task that is just as complex. As we said, P can only perceive and decode meta-interaction through his own usual contextual patterns. As a result of his personal structure and organisation and namely because of his contextual closure, P will tend to interact with T according to the patterns determined by his usual emotional and relational codification. As a general rule, P will tend not to acknowledge the perturbation or to "resist" it by reading it or reacting to it in his own codified ways. Consequently, T's actions will continuously risk being taken as proof of P's expectations, as yet another "quod erat demonstrandum." The second goal of meta-interaction, again instrumental and subordinate to the interactive position, is therefore that of favouring the disarticulation of this redundant contextualisation and of progressively weakening P's "system of proofs" so as to achieve a change in his contexts and a gradual broadening of his relational competences. P will tend to build his context through his relatively fixed ways and rules, so it can be inferred that the second task of a meta-interactive position should be that of facilitating a gradual modification of P's rules of contextualisation.
The posing of the problem of interaction within the global therapeutic frame by setting the two different domains of interaction and meta-interaction offers advantages both at a strategic - that is to say epistemological, theoretical and conceptual - level, and at a more concrete one.
Concerning the first point, it may be enough to say that the redefinition of the formal object permits a more coherent epistemological approach, that above all is able to maintain both the necessary methodological reduction with an "external" point of view and the unavoidable "internal" epistemology of the subsystems P and T and of the entire T-P system. This enables us to retain all we generally mean by transference-countertransference, certainly the unrenounceable wealth of the psychoanalytical method. We actually free it from any naturalistic residues, by placing it in a logical and methodological frame that is adequate to the epistemology of the subjectual systems and to the entire dual system, itself the fundamental consequence of an "internal" point of view and of the necessary autonomy of the systems. In a broader scale, a distinction between "internal" and "external" points of view unties the Gordian knot which, against any epistemological warning, has been keeping inextricably bound together technical-clinical work, research and theoretical development within the session. That way, also in psychoanalysis, the field of research and that of clinical practice can find a more logical articulation.
Above and beyond a strategic level, which cannot be squeezed into a few lines, what has so far been proposed offers a series of advantages in some relating fields:
a) First of all, the distinction between interaction and meta-interaction allows us to set limits to the all-inclusive overgrowth of the technical field. From many sides the complaint has come that this had been expanding over the past decades to include the entirety of therapeutic action. Here technique is cut down to its role of sheer procedure.
b) This distinction defines a realm which was already separate from that of technique in the Freudian vision. This need has often been signalled in the past three decades, although often in the form of blurred conceptualisations such as "real relation", "therapeutic alliance", "holding", "empathy" and such. The notion of interaction can easily include all of these instances, at the same time avoiding the indefinibility of these notions and opening a "window of variables", through which a scientist-observer could look by means of empirically falsifiable conjectures.
c) This position implies a circular articulation between the "attachment" factors and the technical and cognitive ones. This, in spite of all, is consistent with tradition, which, starting with the Freudian conception of positive transference, has always attributed a relevance to attachment factors. This notwithstanding, here the main role is given to the interactive factors rather than cognitive and technical ones. Indeed, a lot of research describes, differently from tradition, what actually takes place in the consultation room and indicates factors of change as pertaining to the field of the experienced relation.
d) To separate the field of technique from that of interaction may lead to a reformulation of method in more operational terms with a subsequent development of more formal tools for contextual analysis concerning the various chapters of technique, such as transference(18) or resistance. This may also help in reducing the bit that is left to common sense.
e) Albeit apparently elusive, interaction can indeed be chosen as the focal point of research and seen as an actual "window of variables. It can be explored at various levels from both a descriptive and classificatory point of view and from one which is closer to actual research on the causes and factors of change. According to a general idea that the causes of change should generally be positively correlated to the intensity of structural coupling, it is also possible to imagine that change can be differently correlated with different types of coupling. For example, it is admissible to imagine a strong structural coupling that strengthen P's anticipations instead of weakening them. At this point we have the task of classifying and describing the possible configurations not simply of the basis of psychopathology - which generally stops with a description of a subsystem P- but on the basis of the configurations of the global P\T system.
f) Finally, as far as factors affecting change are concerned, research can be carried out at different levels: from very "high" hypotheses, such as Luborsky's CCRT, or from a "lower" one relating to graded unconscious expectation systems such as the ones approached by San Francisco Psychotherapy Research Group (WEISS, SAMSON , 1986) with the Plan Formulation Method or variables relating to a processing defined at a more elementary level, according to hypothesis analysed by, for example, Seganti (SEGANTI, 1995), or also from more clearly micro ones, relating to conjectures which could easily be inferred from Damasio's model (DAMASIO, 1994).

This discussion relies on epistemological and conceptual assumptions (relating to the complex structure of biosystems), which it is not possible to discuss here. It shall be enough to say that it relies on a conception of knowledge as a complex process, which refers problem-solving to the entire organism and not on the practice of partial functions, be they intellectual or emotional. Accordingly, knowledge cannot be reduced to a "factor of change": indeed, knowledge is change and change is knowledge. To change, therefore, is to know; being aware of this knowledge, even in the form of insight, equals meta-knowledge. It follows that there is little sense in the contraposition of cognitive factors and attachment factors, or in the compromissary juxtaposition between knowledge and attachment. Attachment or, more precisely, structural coupling, is not a factor of change: it is the inevitable matrix, the ecological niche where " change" becomes necessary and possible, in the same way that it may be necessary and possible (due to the boundaries of their structure and to the potentials of their nature) for seagulls to modify their originary habit of sea fishing and turn it into "fishing for waste", or for starlings to become inurbated and "acquire" the urban environment as the most suitable for spending the night.
"Cognitive" factors and "attachment" factors can artfully be contraposed, separated (or combined) only on the basis of prefixed categories of conceptualisation, that is to say on the basis of the way and of the context used to define the problem-object "change." Freud's representationalist and rationalist mechanicism defined the problem-object change within the terms of his physicalist apparatus, heavy with representations granted by history, so that the economy of investment could be modified through knowledge inferred by the analyst and communicate to the patient. This is what "was supposed to take place" according to theory. What actually took place was that, against all expectations, the patient did not change through knowledge. Resistance and transference came to light, actually a transference of resistance, which theory took in as a detour, as a knot to be untangled before the theorem of knowledge could start. As he was busy perfectioning his epistemic device, Freud unwittingly accepted to live a situation of structural attachment with his patients, where "change" could become necessary and possible, and where, apart from techniques and theories, something was taking place that today we are beginning to understand under a new light.
It is somehow paradoxical (though historically understandable) that the discipline that gave men the meaning and value of childhood experience, of attachment and emotions, the discipline that taught us to take people's "stories" and turn them into the "story" of their "being with" their mothers and fathers, and that even brought the general parameters of all these stories within the "greater story" of philogenesis and evolution - should be based on an intellectualist theory of change and should so painfully struggle to recognise what it has itself promoted.

ASSOUN P-L., (1981) Introduction a l'épistémologie freudienne, Payot, Paris.
ATLAN, H., (1972) La theorie de l'informatiom et l'organisation biologique, Hermann, Paris.
BORDI S., (1995) Lo stato attuale del concetto di neutralità analitica, Rivista di psicoanalisi, XLI, 3, pp. 373-390.
CERUTI M., (1989) La danza che crea, Feltrinelli, Milano.
DAMASIO A. (1994). Descartes' Error. New York: Grosset\Putnam.
DENNETT D. C,. (1991) Conscousness explained. Little, Brown and Company, Boston, London.
EDELMAN G., (1987), Neural darwinism. New York: Basic Books.
EDELMAN G., (1989), Remembered Present, New York: Basic Books.
Von FOERSTER H., (1981), Observing systems, Intersystems Publications, Seaside (Cal.).
FREUD S., (1919), Lines of advance in psycho-analytic therapy , SE, 14.
FRIEDMAN L. (1988) The Anatomy of Psychotherapy, The Analytic Press, Hillsdale, NJ.
GILL M.M., (1994). Psychoanalysis in Transition. Hillsdale, NJ.: Analytic Press.
GREEN A., (1993), Two Discussions of "the inner experiences of the Analyst" and a reponse from T Jacobs. Int. J. Psycho-An., 74, 1131-1136.
HOFFMAN I. Z. (1983). The patient as interpreter of the analyst's experience. Contemp. Psychoan., 19, 389-422.
LUBORSKY L., DIGUER L., LUBORSKY E., SINGER B., DICHTER B. SCHMIDT K. A., (1993) The efficay of dynamic psychotherapy. is it true that "everyone has won and all must have prizes"? in MILLER N., LUBORSKY L., BARBERJ.P., DOCHERTY J (editors), Psychodynamic Treatment Research. A Handbook for Clinical practice, Basic, Books, New York.
HOFFMAN I. Z. (1983). The patient as interpreter of the analyst's experience. Contemp. Psychoan., 19, 389-422.
HOFFMAN I. Z. (1998). Ritual and Spontaneity in the Psychoanalytic Process: A Dialectical-Constructivist View. Hillsdale, NJ: The Analytic Press .
MATURANA H., VARELA F. (1980). The Tree of Knowledge. Boston: New Science Library.
MIGONE P., (1995) Terapia psicoanalitica, Angeli, Milano.
MITCHELL S.A (1988). Relational Concepts in Psychoanalysis, Cambridge, MA: Harvard Univ. Press.
OREMLAND J., (1991) Interpretation and Interaction: Psychoanalisis or Psychoterapy, The analytic Press, Hillsdale, NJ.
PONSI M., FILIPPINI S., (1996) Sull'uso del concetto di interazione, Riv. Di psicoanalisi, XLII, 4, pp. 567-594.
RAPAPORT D. (1967). The Collected Papers of David Rapaport (M.M. Gill Ed.). New York: Basic Books.
RAPAPORT D., (1977) La metodologia scientifica della psicoanalisi (1944-1948), in Il modello concettuale della psicoanalisi, Feltrinelli, Milano.
RENIK O. (1993). Analytic interaction: Conceptualizing technique in light of the analyst's irreducible subjectivity. Psychoanal. Q., 62: 553-574.
RENIK O. (1996). The perils of neutrality. Psychoanal. Q., 65: 495-517.
RENIK O. (1999), "Reactions to 'Observing-Participation, Mutual Enactment, and the New Classical Models,' by Irwin Hirsch," Ricerca Psicoanalitica, Vol. 3.
SCANO G.P., (1982) Il Super-io tra sessualità e socialità, Città Nuova, Milano.
SCANO G.P. (2000). La nozione di interazione e il suo significato per il metodo e la teoria. Psychomedia,
SCANO G. P., MASTROIANNI A.,CADEDDU G., (1995), Psicoterapia psicoanalitica verso una rifondazione, Angeli, Milano.
SCHWABER E.A , (1995),Toward a Definition of the term and Concept of interaction. Int. J. Pshycho-An., 76, 557-564.
SEGANTI A., (1995), La memoria sensoriale delle relazioni, Boringhieri, Torino.
STOLOROW R.D., ATWOOD G.E., . (1992). Contexts of Being: The Intersubjective Foundations of Psychological Life. Hillsdale, NJ: The Analytic Press.
THOMä H., KäCHELE H., (1987). Psychoanalytic Practic, I. Principles. Springer -Verlag, Berlin- Heidelberg
WEISS J., SAMSON H. & THE MOUNT ZION PSYCHOTHERAPY RESEARCH GROUP, (1986) The psychoanalytic Process: Theory, Clinical Observation and Empirical Research, Guilford, New York.


1 About the various psychoanalytical "methods", Rapaport wrote: "I would think that the first specifically psychoanalytical method in terms of importance is the psychoanalytical constellation: a stable relationship between two persons. In the future I shall refer to it as 'interpersonal relationship'. This seems to be the most specific characteristic of the psychoanalytical method" (My translation.). D. RAPAPORT, La metodologia scientifica della psicoanalisi (1944-1948), in Il modello concettuale della psicoanalisi, Feltrinelli, Milano, 1977, p. 112.

2 In 1948 Rapaport realised that if you read interpersonal relationships in the light of the assumption (logically necessary for methodology and for metapsychology) of psychic continuity "...a contradiction arises that is difficult to rectify. The contradiction lies in the fact that this bipolarity is not truly such since the analyst is part of the continuum. It is a difficult situation that causes lots of complexities" (Ib., pp, 124-125, note 9). The contradiction stems from the fact that metapsychology can only treat objects as the intrapsychic representations of objects.

3 The double panel dedicated to interaction by the American Psychoanalytic Association (1991-92) is a clear sign of a shift in focus. Reports of the two debates were published on the Journal of the American psychoanalytic Association: the first by D. M Hurst (JAPA, 43,2, pp.521-537, 1995;), the second by S.D. Purcell, JAPA. 43, 2, 1995, pp. 539-551.

4 In those years, even M. Gill - who had co-authored Rapaport's essay on points of view and together with him had thought about, acquiring it, the one on topography - turned to clinical practice and used it as his starting point for strong revisions to the notion of transference, leading him to a constructivist approach and to interaction. He approached clinic and technique aiming at theory. Instead, his words were understood and measured by many on the basis of codified technique.

5 For an outline of the formal structure of Freudian theory we must refer to D. Rapaport's, specifically to his concise and unmatched The Structure of Psychoanalytic Theory: a systematising attempt, Psychol-Issues, 6, 1960. A careful analysis of the conceptual grounds and of the structure of metapsychology in the epistemological, philosophical, conceptual and scientific setting of the late 19th century, has been outlined by ASSOUN P-L., Introduction a l'épistémologie freudienne, Payot, Paris, 1981. I have carried out historical-critical analysis of the development of Freudian theory in G.P. SCANO, Il Super-io tra sessualità e socialità, Città Nuova, Milano, 1982.

6 Cfr. ASSOUN P-L., op.cit.

7 For example those of Edelmann (1987, 1989), Damasio (1994), Dennett (1991). This paper rests specifically on A Damasio's theory.

8 Friedman gave a careful description of the historical decline of the conflict between the two faces of psychoanalysis, underlining the critical points in the debate and their theoretical constructs, thus showing the actual relevance of this antinomy from a clinical point of view. FRIEDMAN L. The Anatomy of Psychotherapy, The Analytic Press, Hillsdale, NJ, 1988.

9 As Migone recently pointed out (1995, pp.102-103), "in re-reading Freud's work we can evidentiate two main factors that characterised the theory of cure in psychoanalysis: one could be defined as an intellectual, cognitive understanding, as exploration, insight, interpretation and so on, while the other refers to the affectionate bond with the analyst, to the emotional relation and it has been called in different ways, including positive transference, attachment, etc.."

10 Of course this does not deny that a theory of mind would still be necessary.

11 Concerning the notion of "complexity", we must turn to the works of Von Foerster H., Morin E., Atlan H., Varela F., Prigogine I.,...

12 The psychoanalytical vocabulary has two partially overlapping terms for the same field. It is the opposition between symmetry and asymmetry, which generally holds a technical descriptive value concerning the relational state of the two actors. For this reason we preferred to introduce two new terms.

13 The notion of "structural coupling" (Maturana & Varela, 1980) is used in biology to indicate the relation of structural congruency between an organism and the environment in which its ontogenesis takes place, or that between a cell and a multicellular unit, or that which takes place when two or more organisms, during their ontogenesis, constitute a network of recurrent and stable interactions, which, then, form systems or units of a superior, specifically "social" level.

14 This vision rests on A. Damasio's theory (1994) of the semantic marker and more generally on his broader conception of the role of emotions and feeling within the structure and functioning of the mind.

15 This has repeatedly been underlined by Weiss' and Samson's hypothesis and research (1986). In Italy, this was done by the hypotheses underlining A. Seganti's empirical research (1995).

16 An attempt to coherently describe the interactional space - where some roles of interaction can also be found - is offered by A. Seganti, starting from the threat of coherently organising one's self and of the negative anticipations that "...scarcely perceived by the subject's conscience but strongly perceived et a sensory level (...) take on an unavoidably prioritary form and (...) push the patient (...) against his will, to reproduce negative situations in new relationships". Of course this is also the case in the therapeutic setting so that the patient is in the contradictory position of going into analysis to change that kind of situation which he is disappointed to realise to be repeating. In the meantime, the therapist who cannot avoid participating, can only approach the negative situation which keeps coming up by acknowledging his own role in the creation of such a situation. A(nalyst) identifies three possible relational outcomes called "reciprocity, exclusion, inclusion". SEGANTI A., op. cit., pp. 96-107 (passim).

17 This hypothesis is the one that is used most often to explain the paradox of equivalence, that is to say, the fact that the effectiveness of different psychotherapies does not seem to be related to their specific techniques. However, there is no univocal interpretation of this data. Luborsky and AA. list eight possible reasons to explain the paradox of equivalence. L. LUBORSKY, L. DIGUER, E. LUBORSKY, B. SINGER, B. DICHTER, K. A. SCHMIDT, ( 1993).

18 Concerning transference, I made an attempt in that direction in my "Phlogiston and the Transferential 'Thing'", in press.