Introduction.

The construct of Therapeutic Alliance (T.A.) shows, in its many definitions, problematic theoretical divergences. It is surprising to see that, despite these divergences, the Therapeutic Alliance is usually accepted as the unspecific factor in “all” kinds of psychotherapy and is universally considered the most important predictor of the outcome of psychotherapy.
Let us see the “key words” used in authoritatively defining the Therapeutic Alliance. Clara Zetzel, in 1958, was the first to use the expression T.A. and she described it as “mature sharing of the reality between patient and therapist in psychoanalysis1” (Lingiardi, 2002). Luborsky (1976), in the type 2 Helping Alliance, talks about the “consciousness of the patient and the therapist that they are involved in a common process”. Louise Gaston subdivides the T.A. into four aspects: 1- Patient working capacity, 2- Patient commitment, 3- Working strategy consensus, 4- Therapist understanding and involvement (Lingiardi, 2002). It is interesting to see that the first two aspects are related to the patient, the last to the therapist and the third to a consensus on strategies that can only refer to both the protagonists in the relationship. Bordin (1979) defines the T.A. as the reciprocal agreement on the Goals of change that will be sought with psychotherapy, on the Tasks that can help to attain these goals and on the establishment of Bonds with the role of maintaining the collaboration between patient and psychotherapist. Aaron T. Beck, the well-known cognitive psychotherapist (quoted by Lingiardi, 2003), talks about a joint effort to reach a congruence between the patient’s expectations about therapeutic goals and those of the therapist. We will stop there for the moment, seeing that Bordin’s definition inspired Horvath and Greenberg (1994) in their construction of a tool to assess the T.A., the Working Alliance Inventory (WAI), one of the best known psychometric tools for finding the level of T.A. in psychotherapy. The analysis of the factors that make up the WAI, using specific tools that will shortly be described, is the objective of the research presented in this article.
Let us look at the key words in the definitions just given: sharing (in Italian, ‘condivisione’), consciousness, common, consensus, [reciprocal agreement]; collaboration; congruence. Isn’t it interesting? To talk about the T.A. all the authors mentioned, but certainly also others in a deeper analysis of the endless literature in the field, use words in which the first part of the composite term is the Latin ‘cum’, meaning with or together: cum divido; cum sapio; cum munus; cum sensus; cum laboro; cum gruo. The implicit sense in the etymology of the words is “doing something together”.  The words used therefore imply the notion of reciprocity. Let us now think of the origin of the construct of T.A.: for a long time there was controversy over whether or not it was a different phenomenon from transference, a central aspect of the theory of psychoanalytic technique. But it is important to notice that transference is a process by definition individual, concerning the patient as an individual. See for instance the definition of transference given by Laplanche and Pontalis (1967) in their “Vocabulaire de la psychanalyse” (translated into Italian with some exaggeration as “Encyclopaedia of psychoanalysis”): “The term transference indicates, in psychoanalysis, the process by which unconscious desires are actualised on certain objects in the context of a particular relationship established with them and above all in the domain of a relationship of analysis”.
This involves a repetition of childhood prototypes that is experienced with a strong sense of present reality.
It is mainly the transference of the treatment that is called transference by psychoanalysts, with no other specification.
Transference is traditionally recognised as the terrain of psychoanalytic treatment, its design, its modalities, its interpretation and its conclusion.
The definition cited above, too, refers to single individuals, the patient and the psychotherapist. Moreover, when T.A. is talked about, there are systematically aspects concerning the relationship between the two. It is in this regard that compound words are mentioned, where one of the components is the Latin term “cum”. Let us think about the fact that the notion of T.A. emerged from the theoretical and technical context of psychoanalysis. It can be asked, rightly we think, why, to define this notion, which is related (unlike transference, and also countertransference) to a relationship and not to two individuals, are “rational” aspects resorted to, such as agreement and reciprocity, neglecting the more important  psychoanalytic dimension, the mind’s unconscious mode of working. The answer, we believe, is simple and astonishing at the same time: classical psychoanalysis has used the notion of the unconscious to talk about single individuals, in their inner world, not about relationships. The experience of psychoanalysis, on the other hand, is based on the relational process between patient and psychoanalyst, both in the agreement on the setting2 and in the dynamics of the analytic relationship. The individualist conception of the psychic disorder and of psychotherapy leads to the relational aspects being considered “outside” the psychological theory underlying the psychotherapeutic intervention itself. But these relational dynamics, not considered in the therapeutic process due to the lack of a theory of the technique concerning the relationship3 and not the individual, together justify the interest in the Therapeutic Alliance. In other words, the individual’s psychopathology is made the object of psychotherapy and, to make sense of the treatment of this psychopathology related to the single individual, individualistic theories of the psychotherapy intervention are needed; the relationship between psychotherapist and patient, on the other hand, supports the likelihood that the psychotherapy will take place, and that it will last. Hence the importance of the relationship, of its reciprocal nature, of the agreement that “just” govern the identification of aims and methods to achieve them. The T.A., therefore, is offered as the substitute construct of a theory analysing the relationship as the object of psychotherapy, not as the integrative condition for its realisation.
We have presented a psychoanalytic theory of the relationship, grounded on the construct of collusion: the emotional symbolisation of the context by those who share a specific context is the basis of the relationship in its symbolic components. With the Analysis of the Demand we have also proposed a model of the relationship, specific to the psychoanalytically-oriented psychotherapy experience. In a psychotherapeutic relationship considered as the outcome of the collusive process, it is not necessary to postulate the T.A.: the symbolisation of relating organises the therapeutic relationship. With the failures of collusion induced by the transformation of the relationship, the dynamics of collusion can help to understand the vicissitudes of the therapy in its various facets and in its process.
It is not our intention to examine these statements further, since they have been studied in other works. What we want to present in this research is a “deconstruction” of the construct of T.A., as it emerges in Horvath & Greenberg’s WAI, is it is used with a specific methodology based on the Emotional Textual Analysis.

The measurement of the Therapeutic Alliance.

We have mentioned the deep contradiction that can be found within a construct, that of T.A., often used to talk about different things in psychotherapy, from the “intellectual collaboration of the analysed” mentioned by Freud in his Recommendations to Physicians practising Psychoanalysis from 1912, through to the breakdowns in patient – therapist collaboration in the T.A., presented by Safran & Muran and taken up by Colli & Lingiardi.
It is interesting to see that in the psychology of the  Eighties and Nineties, especially in the United States and Canada, the need was felt to use suitable scales, inventories, or other psychometric tools to measure the constructs whose theoretical directions were being advocated. The construction of a tool to measure a theoretical construct, on the other hand, has the undeniable power to “reify” the construct itself, making it become an a-critically adopted “thing”, insofar as it is measurable and can therefore be translated into a number, a measurement that can be processed from the statistical point of view. The examples of this are numerous: from neuroticism to empowerment, from alexytimia to transference, from attachment styles to the T.A. We would like to draw attention to the difference between constructs such as Herman Witkin’s field dependence-independence, elaborated using the empirical data collected in the study of perception and organised to explain the data itself, and the psychometric tools we are talking about: measurement tools proposed to assign a numeric value to constructs emerging from clinical experience and built with the intention of giving a measurement of events that can be phenomenonically linked to similar dimensions. In the first case the construct unifies the empirical data, in the second the tool tries to give a measurable status to a construct defined conceptually, often with contradictions and different points of view. The reification of a conflictual construct, through a measurement tool, puts an end to the debate about the meaning of what is measured. This can be a real problem in the scientific debate of the construct, given that data that is often contradictory places in doubt not the tool, but the application of the construct. Think, purely as an example, of the research on the relation between the alliance and the seriousness of the disorder: some authors find that the more serious the disorder, the weaker the alliance; others have found no difference in the alliance, in relation to the seriousness of the problem being treated. The a-critical acceptance and the a-critical use of such tools can lead to confusion and redundancy of data, often contradictory, in research in the sector involved.
The Working Alliance Inventory, by Horvath & Greenberg, is based on Bordin’s definition of T.A.; this author considers the T.A. as a mutual agreement between patient and therapist on goals, tasks and bonds, as has already been indicated. The Italian version of the WAI measurement tool is composed of 36 items that can be subdivided into three scales: the Goal scale with 12 items (6 in affirmative form and 6 negative), Task scale with 12 items (7 in affirmative form and 5 negative), Bond scale (9 affirmative and 3 negative). There is also a short version of 12 items.
There are three versions of the WAI: one based on the patient’s point of view (WAI-P), the second on the therapist’s point of view (WAI-T) and the third representing the point of view of an external observer of the therapy (WAI-O). There is a great deal of research in which the application of the WAI has enabled the T.A. to be measured, on various experimental levels. It is through this research that the T.A. has proven to be the construct best able to predict the outcome of psychotherapy, in turn measured with different methods.

The research methodology.

We know that an experience of the unconscious “in the pure state” is impossible: as Matte Blanco says, we would plunge into a state of non-existence. This is due to the fact that “existing”, i.e. the experience of being in time and space, entails heterogeneous, dividing thought. But the emotional symbolisation of the different aspects of reality is always present in our mind, and grounds the collusive social relationship. This also occurs in psychotherapy. Psychoanalytic psychotherapy, like other human experiences, is made up of components based on heterogeneous dividing thought, which establishes relations and therefore is able to separate and distinguish the objects of reality, and of components based on homogeneous individing thought; the latter determine the emotional dynamics of experience. The first components come from the setting (time and duration of the sessions; payment; arrangement of the session with the patient lying on the couch and the analyst seated behind, out of the patient’s sight: this arrangement is designed to foster a relationship between patient and therapist where the stimulus from the exchange of glances is reduced to a minimum and where the patient’s associations are based on the prevailing emotional symbolisation of the relationship); the second components come from the free associations and the emotional events that mark the analytic relationship. As we can see, agreements on the aims of the psychotherapy  or on the bond are not envisaged; the setting and the fact that it is shared is the only element organising and making possible the psychoanalytic experience. Any transgression of the setting, being necessarily on the level of enacting, cannot be analysed and may lead to the breakdown of the therapeutic relationship. That is why it is important for the patient to be informed of the conditions of work required by psychoanalysis. What happens in psychoanalysis only acquires a meaning by being analysed; analysing means thinking of the emotions aroused by the situation of the setting, in the psychoanalytic relationship. WAI, on the other hand, calls for the use of dividing thought to assess some aspects which, in our vision of psychotherapy, are related to the unconscious dynamics of the psychotherapeutic relationship. It has been said that Bordin’s definition of T.A. sought to be “pan-theoretical”, in other words, designed to show the components of the alliance in all forms of psychotherapy: psychoanalytic, cognitivist, Rogersian. We think that this expanded vision of the alliance, and therefore of the relation between psychotherapist and patient, has been to the detriment of the emotional dynamics working in psychotherapy, characterising the psychoanalytic vision of treatment. This means that, in the formulation of the WAI items, there is the attempt to communicate on a cognitive level, that which actually belongs to emotional dynamics; therefore the attempt to force everything related to emotional processes that cannot be confined to the three categories G (goals), T (Tasks) and B (Bonds), into an assessment based on the meaning of the statements in the WAI.
Let us look for instance at WAI item 4, presented below: “Patient and therapist are confused about what they are trying to do in therapy”. We have said that in the psychoanalytic relationship, patient and psychotherapist have a relationship originating in the collusive symbolisation of the context coming from the setting of the treatment. As we have suggested elsewhere, collusive symbolisation is polysemic, and therefore highly undifferentiated, ambiguous, marked by diverse emotions and by the constant process of differentiation-nondifferentiation of the emotions themselves. This means that the psychoanalytic relationship is confused. It is not the patient or the therapist that is confused, but the relationship between them, a relationship that originates in the collusion process. If the  psychoanalytic relationship is seen in this light, the confusion if highly likely. On the other hand, if by confusion we mean, in  psychoanalytic theory, the entrance of the mind’s unconscious mode of working into the transactions in the relationship, how can an “ordinal” score be given to the confusion? This objection does not concern only the items under discussion, but all the WAI items.
We will now present the short form WAI-O in the Italian version.

We have shown in [ ] the attribution of the single items to the three scales: Goals [G], Tasks [T], Bonds [B].
A Therapeutic Alliance assessment is obtained by giving a score to the single items. This score, in the methodology underlying the measurement tool, involves a “cognitive” reading of the individual statements, based on the shared sense of each utterance. Our hypothesis, by contrast, is that the score assigned is affected by the affective symbolisation evoked by each statement. This emotional reading of the items, on the other hand, can reorganise the sense of the Therapeutic Alliance, which might not correspond to its subdivision into the three scales, as envisaged by the authors of the measurement tool. In the statements, in fact, there is the systematic recurrence of “dense words”4, such as: improve, trust, usefulness, problems, like the other, confusion, ability, effort, need, agree, respect, agreement, mutual, real, understanding, change, solve, right. These “polysemic” words, can regulate an emotional representation of the patient-therapist relationship, which is different from that emerging from the three scales; they can give specific information about the relationship, which cannot be provided by the descriptive schema of the T.A. given by Bordin and then transferred into the WAI by Horvath & Greenberg.
Hence the hypothesis of this research: to replace the assessment via scoring of the single items of the short form WAI-O5, with the verbal associations evoked by each individual item; these associations are collected from psychotherapists who associate while thinking of a specific therapeutic relationship they have experienced with a particular patient, chosen amongst those in therapy. These associations replace the assessments expressed via 1-7 scores on each item, and can give us information on the emotional symbolisations evoked by the statements in the items.  This was done by asking for the collaboration of a group of 28 psychotherapists of various theoretical orientations6.
Here are the characteristics of the 28 psychotherapists interviewed:

We asked each psychotherapist to write a short report to present a patient chosen among those currently in therapy. This was in order to focus the psychotherapist’s thoughts on a specific patient. Later, we asked them to free associate on the WAI items, presented one at a time, with reference to the patient chosen and described in the report. We recorded the associations on the 12 items. We subjected the whole text of the associations including the associations of all the psychotherapists, for all the WAI items) to Emotional Textual Analysis, keeping the subdivision by item and by psychotherapist as illustrative variables.
The hypothesis of the research envisages that the analysis of the clusters of dense words present in the psychotherapists’ associations on the 12 WAI items will be reorganised in different structures from the three scales envisaged in Bordin’s theory. The analysis of this expected reorganisation will be able to give us information on the T.A construct, arranging the components in categories closer to the distinctions proposed among the “organisational” components of the therapy and the emotional dynamics of the same.

Results:

For the AET we used Max Reinert’s programme Alceste (Analyse des Léxèmes Cooccurrents dans les Enoncés Simples d'un Texte). In accordance with the ETA method, we intervened on A2 DICO (the dictionary of verbal forms, generated by the programme) accepting only the “dense” words in the dictionary, as the only forms to process, and marking the single “non dense” forms with the letter “r” (rejected) so as to exclude them from the analysis. This operation requires long slow work and for the single verbal forms in the dictionary, a competent choice by the researcher. On the other hand, it enables the analyses to be made, consistent with the AET theoretical hypotheses, which consider the clusters of dense words present in the text as indicators of the different collusive process running through the test itself.
The AET analysed 318 dense words, while 838 verbal forms were classified as “non dense”. We wish to stress this fact, to show that some text analyses can work on all the verbal forms present in the text: in this case there are reorganisations of words that in fact reproduce what is already in the text without any additional information based on psychological theory. It is an analysis of the text requiring little work, but in the end it offers little information in terms of the psychological interpretation of the text.
The text was subdivided by Alceste into 653 elementary units of context (u.c.e.). For the analysis 349 u.c.e. were used, that is, 53.45% of those generated by the computer programme. When the proportion of the u.c.e. processed is over half those generated by Alceste using the text under analysis, then the latter is a good representation of the text that is to be analysed.
Processing the dense words gave rise to four clusters, distributed as follows in the factorial space (in the squares the dense words shown are those characterising the cluster and the falling value of χ², to show the greater or lesser centrality of the single words within the cluster):

Lastly, let us consider the distribution of WAI items in the four clusters.
Items present in cluster 1:
5 - The patient has confidence in the theapist’s ability to help him [B]
8 - Patient and therapist agree on the important things the patient needs to work on [T]

Items present in cluster 2:
1 - Patient and therapist agree on the steps to take to improve the patient’s situation [T]
9 - Patient and therapist have constructed a relationship of mutual trust [B]

Items present in cluster 3:
2 - Patient and therapist are confident about the usefulness of their therapeutic activity (for example, the patient manages to see his problems in a different way) [T]
3 - The patient thinks the therapist likes him and the therapist thinks the patient likes him [B]
4 –Patient and therapist are confused about what they are doing in therapy [G]
7 - The patient feels that the therapist respects him as a person [B]

Items present in cluster 4:
10 - Patient and therapist have different ideas on what the patient’s real problems are. [G]
11 - Patient and therapist have established a good level of mutual understanding on the kind of change that would be right for the patient [G]
12 - The patient believes the path taken to solve his problems is the right one [T]

Item 6, Patient and therapist are engaged in a common effort to reach the agreed goals [G], is not part of any cluster.

One initial comment, based on the data obtained: the relation between the clusters of dense words emerging with AET and the three WAI scales, foreseen by their creators, does not exist. The items belonging to the three scales are randomly distributed in the four clusters obtained with AET.
Let us now analyse the emotional dynamics, as it emerges in the single clusters and in their relation in the factorial space. On this point, we show the values of the cluster-factor relations:

Now we will present the analysis of the single clusters and their relation in the factorial space generated by Alceste.
For the analysis of the dense words that give rise to the individual clusters, we use the methodology indicated in Carli & Paniccia, 2002: the sequence of encounters of dense words is examined (the first dense word with the second, the first and the second with the third and so on); the analysis is based on the progressive reduction of polysemy characterising the dense words, in their encounters. It is from this polysemy that one can identify the collusive sense of the cluster. The position of the different clusters in the factorial space will allow a second level of analysis to be carried out, based on the relation between them.

On the first factor there is a contrast between cluster 1 and cluster 3.

Cluster 1.
The first instance of dense words is: ‘share’ – ‘therapy’. From the beginning of the dense words in the cluster, there is a clear connotation of psychotherapy  as sharing (in Italian ‘condividere’) (from the Latin cum divido meaning split in two or share with the other person). The third dense word is ‘construct’ (from the Latin cum struo, meaning put one thing with another, build together). Then come ‘process’ and ‘exploration’, which indicate proceeding ahead trying to discover: this research is done together. And then there is the word that explicitly brings out what was present in the prefix cum in the previous dense words: ‘together’.
Psychotherapy  is experienced as a process and as a relationship, with ‘together’ being the dense word that indicates the emotional sense of the cluster: there is sharing, there is constructing, complexities are explored; patient and psychotherapist do these things together. The process involves exploration, so there is a hint of venturing into the unknown where being together, sharing the experience, attenuates the risk of this exploration and allows for construction, increasing the complexity (the next dense word) of the relationship itself.
This cluster underlines the function of the relationship in psychotherapy, and the collusive dynamic takes on the aspect of proceeding to construct meaning.
Let us look at the dense words in the WAI items which, among the illustrative variables, are found in cluster 1.
Item 5 stresses the patient’s critical dependence on the therapist (trust in his/her abilities), although this is all in the sense of helping, and therefore of relational models without technique.
Item 8 envisages an agreement, albeit regarding work that the patient “must” do. So again there is the patient’s dependence on the therapist.
But as we have seen, this dependence has critical connotations: the trust is in the therapist’s abilities, not in the therapist tout court; the patient “has to” work on important things, about which, it must be stressed, he is in agreement with the therapist, and therefore participates in defining them. 
The psychotherapists’ associations seem to expand and develop this critical aspect of the patient’s dependent attitude, in a sort of  symmetrisation of the dependency, expressed in exploring together, sharing and constructing. The patient’s dependency, characterised in the WAI items by a critical attitude, is transformed into reciprocity of intentions and actions in the associations.
The WAI items seem to suggest stress on the collusive process organising the sharing of the therapeutic experience.
On the opposite side of the first factor we find cluster 3.
Here there is a contrast between dimensions indicating cognitive assessments, such as capacity and respect, as opposed to problematic emotional dimensions such as  limitation, depression, frustration, and attempt. Whereas cluster 1 shows a symmetrical, gratifying reciprocity, here the relation seems to be based on a profound a-symmetry: on the one hand the patient evaluates the therapist’s ability and feels the therapist’s respect for him; on the other, there emerges the frustration from the experience of a limitation due to the therapist’s dominance, and the resulting depression; it seems that all this may derive from the positive evaluation that helps to place the psychotherapist in a dominant position, being based on a-symmetrical power. The attempt to get out of this asymmetrical dimension seems to be related to the request for greater humanity on the part of the other person, the therapist. It seems to mean that it is not enough to evaluate the other person as capable and to obtain his/her respect in order to avoid the feeling of a limitation and the frustration of the subsequent depression. What is needed is a humanity that brings symmetry to what is made inevitably asymmetrical in a relationship based on evaluation.
 
Looking at the WAI items one might ask where this experience of frustration and asymmetry comes from. It would seem that “pleasing the other person” is a problem constraint, a seductive dynamic involving reciprocal dependency between patient and psychotherapist. Actually, if the aim is to gain the approval of the other person, there  emerges some confusion on what one is doing in therapy; the desire to please the other person constrains one in a dimension of impotence and reciprocal seduction; pleasing the other person seems to be in profound contradiction with the psychotherapeutic situation, a therapy where the usefulness of what is done, and therefore of change, is related only to the patient. The items denote explicit asymmetry (items 2 and 7, where it is the patient that looks at his problems in a different way and it is again the patient that receives the therapist’s respect, without being reciprocated); where this asymmetry is lacking, one feels confused, in an attempt to symmetrise the seduction. The desire to please the other person seems to generate confusion and consequently to increase the asymmetry in the relationship.

On the first factor, we therefore find the contrast between two collusive processes: one seeking sharing and construction, the other seeking mutual liking and asymmetry in this esteem and in the outcomes of the psychotherapy. There is a symmetrical collusion and an asymmetrical collusion, with very different outcomes: the patient “has to” deserve the therapist’s esteem, he evaluates his/her capability, both are in an position of reciprocal seduction; this all leads to a depressive frustration deriving from the feeling of limitation and confusion that the patient (but also the therapist) experiences in this collusive dynamic.

On the second factor there is again a contrast between cluster 1 and cluster 4.
Cluster 4, therefore, may show a different collusive modality from that already shown in cluster 1.

Cluster 4.
This is the cluster where the dense words of the WAI items are repeated: right, real, solve, change. This repetitive aspect can be interpreted as a reason to trigger a reaction: the desire to enact the emotions is presented as the contrast to the prescription about what is “right for the patient” and about his “real problems”. It therefore seems that the prescriptive component, inherent to the WAI items, entails an infantile fantasy of enacting a reaction in the associations of our psychotherapists. This tendency to react is expressed in dense words such as anger, drugs, terrible, at risk, failure. The culture of the cluster seems to be decidedly opposed to the aspects of “reality” indicated in the items: having different ideas on the patient’s real problems, having reciprocal understanding on the changes that are right for the patient, believing that the path taken is the right one.

Remember that the items correlating with this cluster talk about the patient’s “real problems”, about the “right way” to solve the patient’s problems and about a “good level of mutual understanding”. Asymmetry is not related to the relationship between the patient (dependent) and the psychotherapist (possessing power) as in cluster 1 and, even more, in 2; here, “objective” criteria like reality and rightness (justice) are expected to be the parameters the patient has to face up to. The desire to enact emotions therefore appears to be the way to escape from the dominance of the real reality, seemingly to reconfirm the priority of one’s emotionality. In other words, what emerges is the emotional reaction to a therapy presented as seeking to correct a deficit: a deficit objectively defined as a “real problem” for which the patient has to take the “right path”. The threat of the “objective” diagnosis, implicit in the WAI items, seems to lead the associations towards a profound emotional reaction and towards processes of anger and destruction. And the opposite may be true: where there are emotional enactments, where there is anger, failure, drugs, it is important for the therapist to establish a socially accepted way of returning to order. In both cases, there is a clear conflict between the call to order and the emotional reaction.

Cluster 2 is placed on the third factor.
This cluster identifies psychotherapy as development, in contrast to a psychotherapy that seeks to correct a deficit according to realistic parameters, as in cluster 4, described above.
This cluster is characterised by many verbs: develop, grow, rethink, examine in depth, discover, discuss, get involved. Psychotherapy is seen as a relationship of consultancy,  where there is maximum symmetry between patient and psychotherapist. This symmetry leaves space for doubt and discussion, where the consultancy role of the psychotherapist is placed at the service of the development goal, present in the patient and in the relationship.

The statement from the WAI items with the greatest emotional weight in this collusive dynamic aiming at development is: “mutual trust” and “in agreement”. The aim is the “improvement” of the patient’s situation, without mentioning problems, injunctions or unrealistic parameters. Subjectivity is saved, as the element that places the two protagonists of the relationship in a symmetrical position.

Lastly, remember that item 6 of the WAI (Patient and therapist are involved in a common effort to reach agreed goals) is not included in any cluster. In the comparative analysis of the data it will be possible to formulate some hypotheses on this failure to include the item in any of the four clusters.

An emotional revisitation of the notion of Therapeutic  Alliance.

What kind of Therapeutic Alliance emerges from the dense words present in the associations with the WAI items made by the 28 therapists participating in our study?
The stimulus provided by the WAI seems to evoke, in the first and in the second factor of Emotional Textual Analysis, a shared symbolisation of power in the therapeutic relationship. This means that amongst our therapists, the associations with the WAI items in the clusters of the first and second factors evoke collusive feelings that can be linked to a power relationship between therapist and patient. The alliance is therefore represented here by acceptance of or reaction against the therapist’s power over the patient. Cluster 1, which is contrasted to cluster 3 on the first factor and to cluster 4 on the second factor, seems to be central to this theme. In cluster 1, as we have seen, the collusive dynamic seems to aim at overcoming the patient’s dependency on the therapist. This is suggested by the dense word “together” where the symmetry implicit in the word allows psychotherapy to be represented as exploring, sharing and constructing. It is some specific signals of symmetry present in the WAI items situated in the cluster, that prompt and promote this dynamic in order to make the patient’s dependency on the therapist more symmetrical and shared.
The T.A. is therefore offered as the way to overcome the dependency and the differentiation between patient and therapist, within the symbolic emotional representation of power. “Together”, as we have seen, is the word that continues to keep the two figures of patient and therapist separate, but at the same time unites them in the constructive function, around which collusion is organised in order to represent the therapeutic relationship. A construct, from the Latin: cum struere where the verb struo-ere means place in layers, and therefore build, and cum means together; hence the imprecision of the expression co-construct, often used in clinical psychology to express a meaning which is already completely expressed by the form “construct”.
Let us see which dynamics opposing this collusive process are found on the first and second factor.
The first presents asymmetry in the assessment of the therapeutic process and of the esteem where it is the patient alone who sees his own problems in a different light or feels esteemed by the therapist; in parallel, symmetry is evoked in a sort of reciprocal seduction  which suggests confusion and limits. Hence the depressive frustration and the demand for greater humanity on the part of the psychotherapist.
On the second factor, on the other hand, the asymmetrical power is depersonalised in dimensions of reality and justice or, as Roland Barthes (1957) would say, justness; the reaction seems powerful here, and it is all channelled into the emotional enactment, opposed to this call for order. This order is powerful, not being embodied in the psychotherapist but in the elements of reality and justness that he or she embodies for the patient.
In sum, in the first two factors the important element, the symmetry/asymmetry of power between patient and psychotherapist, seems to induce different collusive processes: when symmetry is overcome, there is the possibility of “together” or, again as Barthes would say, of “togetherness”; if this asymmetry of power is underlined by the emotional stimulus (dense words) inherent to the WAI items, then there are reactions that tend to evoke depressive frustration or to exalt the emotionally enacted reaction.
In the third factor, individualities are overcome and incorporated into the development aims, where there is no longer a distinction between the figure of the patient and that of the psychotherapist. The collusive dynamic here integrates the two protagonists into one goal which no longer belongs to just one and, perhaps, to the other. It is no longer consensual, but shared. In this cluster, collusion involves the symbolic sharing of the process characterising psychotherapy. It is clear that in the collusive process of this cluster, we are not in the construct of Therapeutic Alliance between patient and psychotherapist, but we find the symbolisation of the reaction tending towards development.
It can therefore be said that in the associations of the psychotherapists in our research, the WAI items evoke two orders of collusive representations of the psychotherapeutic relationship:

- a relationship based on the asymmetry of power between patient and psychotherapist7; here the two protagonists of the relationship are oriented to emotionally symbolising the other as a figure endowed with power or as the victim of his own power, and to structuring defensive processes against this emotional symbolisation (in cluster 1 “asymmetry is symmetrised” and working together can be symbolised; in cluster 3, asymmetry causes depressed frustration; in cluster 4 there is the emotional reaction to the asymmetry objectivised in terms of reality and justness). It is important to notice that, in all three clusters, patient and psychotherapist remain quite separate and seem to tend towards confining their collusive dynamic only to symbolising the other person.
- Interest in the relationship8, where the two interlocutors of the Therapeutic  Alliance “disappear” in a relationship focused on and aiming at a “third thing”, namely development.   This is not the development of the patient, but the development of the relationship in a context. And the focus on the relationship was as a symbolic object complete unto itself, not as the result of the agreement/disagreement between the two protagonists, which allows the relationship with the third element, which, it is important to stress, concerns the relationship, not one of the two protagonists. This development is the notion of the relationship that is opposed to the correction of a deficit in the patient alone.
In sum, the research shows that the WAI items, designed to measure the Therapeutic  Alliance, evoke problems of the symmetry–asymmetry of the symbolic power between  patient and therapist; problems that can be acted out in the relationship, through different defensive processes, or that can be overcome through the collusive dynamic aimed to see the relationship and its product, development.
Why does psychotherapy deal with power problems? If we think of psychoanalytic psychotherapy, every “real” power dynamic between psychoanalyst and patient is confined to agreeing on the work setting. In the setting, the fantasies connected to power of one over the other, both in the patient’s and the psychotherapist’s mind, are objects of analysis.   In other words, thought gives a sense to the fantasies, the collusive symbolisations characterising the relationship. Thinking of emotions is the typical, constant  psychoanalytic work. In the  psychoanalytic domain, therefore, the Therapeutic  Alliance, outside of the setting, does not arise as the problem. Nor does the breakdown of the alliance. The patient’s fantasies may certainly be critical of the analyst, just as the content of dreams or the free associations may be, during one or more sessions. The problem lies entirely in the possibility of analysing these fantasies, as well as fantasies of complicity, friendship, sharing of therapeutic goals, and whatever passes through the minds of the patient and psychotherapist. Things are very different in other kinds of psychotherapy, particularly in cognitive psychotherapy. Here the power of the psychotherapist over the patient is the main tool for psychotherapeutic work. If we look at the years when the issues related to the Therapeutic  Alliance were at the peak of their popularity, the seventies and eighties, we can see that it was also the period of greatest diffusion and popularity of cognitivist theory and psychotherapy. It was an era when, as we have already pointed out, there was a marked change in the idea of the mind’s unconscious mode of being as an “alternative” logic to the logic of dividing, heterogeneous thinking; this ended up with the notion of cognitive unconscious in a dubious hybridization at the expense of the real, innovative discovery of Freudian thought. We think cognitivism needs the Therapeutic Alliance, as an alternative to the process of analysing fantasies prompted  by the collusive psychoanalyst-patient relationship, in  psychoanalytic treatment. Cognitivism needs the Therapeutic  Alliance as a sanction of the psychotherapist’s power, an important frame for the operation of psychotherapy seen as deficit correction. Being designed to correct deficits, cognitivist therapy is placed in the domain of the medical relation, but lacks the cultural support that the medical act uses in its practice. We think this power dynamic is well expressed in the WAI items, and that it was explicitly perceived in the associations of the psychotherapists participating in our research.
In the context of the associations evoked by the WAI items, moreover, there is no differentiation between the three scales indicated as goals, tasks and bonds. We feel that this may be linked to the fact that the WAI talks explicitly about goals, tasks and bonds, but within an emotional dynamic that makes these elements part of the relationship of the psychotherapist’s power over the patient. The emotional content of the items, in fact, evokes dynamics of power between the patient and the therapist or the possibility of there being a willingness to relate, and therefore to develop the relationship.
Item 6, which does not match any cluster of dense words present in the associations, seems to be the one that best represents the classical view of the Therapeutic  Alliance: “common effort” and “agreed aims”, but also “patient and therapist”. We think this separation between patient and therapist, contrasting with the dense words “common” and “agreed”, places the cluster in a central, undifferentiated position,  between power dynamics and the relationship tending towards development.
One might ask: with other therapists, different from those who contributed to our research, could there be new clusterisations of dense words associated to the WAI items? We think so. We would probably find new dynamics evoked by the WAI items and new organisations of the construct of T.A.. This would at the same time confirm our hypothesis: the WAI items are ambiguous statements9 that express very well the ambiguity of the construct of Therapeutic  Alliance, not only in its theoretical construction but also in the tool that measures it.

References

Barthes, R. (1957). Mythologies [Mythologies]. Paris: Editions du Seuil. Trad. It. (1974). Miti d’oggi. Torino: Einaudi.

Bordin, E.S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research and Practice, 16, 252-260.

Carli, R. (2007a). Notes on the report. Rivista di Psicologia Clinica, 2, 186-206. Consulted 15 May 2009 at
http://www.rivistadipsicologiaclinica.it/english/number2_07/Carli.htm

Carli, R. (2007b). Punchinello or ‘on ambiguity’. Rivista di Psicologia Clinica, 3, 382-396. Consulted on 15 May 2009 at
http://www.rivistadipsicologiaclinica.it/english/number3_07/Carli.htm

Carli, R. (2008a). The report and de diagnosis. Rivista di Psicologia Clinica, 2, 154-170. Consulted on 20 May 2009 at
http://www.rivistadipsicologiaclinica.it/english/number2_08/Carli.htm

Carli, R. (2008b). Le figure dense in Vittore Carpaccio. Rivista di Psicologia Clinica, 2, 247-256. Consulted on 15 May 2009 at
http://www.rivistadipsicologiaclinica.it/english/number2_08/Carli_b.htm

Carli, R., Grasso, M., & Paniccia, R.M. (2007). La formazione alla psicologia clinica: Pensare emozioni [The trainng to clinical psychology: thinking emotions]. Milano: FrancoAngeli.

Carli, R., & Paniccia R.M. (2002). L’analisi emozionale del testo [Emotional text analysis: un instrument for reading texts and speaking]. Milano: FrancoAngeli.

Carli, R., & Paniccia, R.M. (2003). Analisi della domanda: Teoria e intervento in psicologia clinica Analysis of demand: Theory and technical of psychological clinical intervention]. Bologna: Il Mulino.

Carli, R., & Paniccia, R.M. (2005). Casi clinici: Il resoconto in psicologia clinica [Clinical cases: the report in clinical psychology].. Bologna: Il Mulino.

Codignola, E. (1970). Il vero e il falso: Saggio sulla struttura logica dell’interpretazione psicoanalitica  [The true and the false: Essay on the logical structure of psychoanalytic interpretation]. Torino: Boringhieri.

Freud, S. (1912). Consigli al medico nel trattamento psicoanalitico [Suggestions to medical in psychoanalytic treatment]. OSF, 6, 532-541.

Horvath, A.O., & Greenberg, L.S.(Eds.) (1994). The working alliance: Theory  Research and Practice. New York : John Wiley & Sons.

Laplanche, J., & Pontalis, J.B. (1967). Vocabulaire de la psychanalyse [Dictionary of Psychoanalysis]. Paris: PUF. Ital.trans. (1968). Enciclopedia della psychoanalysis. Bari: Laterza.

Lingiardi, V. (2002). L’alleanza terapuetica: Teoria, ricerca e clinica [Therapeutic Alliance:Theory, Research, Clinical]. Milano: Cortina Editore.

Lingiardi, V., & Colli, A. (2003). Alleanza terapeutica: Rotture e riparazioni [Therapeutic Alliance: Breakings and Repairs]. Quaderni di psicologia, analisi transazionale e scienze umane, 38. Consulted on 20 May 2009 at
http://www.psychomedia.it/cpat/articoli/38-lingiardi-colli.htm

Notes

* Full professor, Faculty of Psychology 1, “Sapienza” University of Rome. Top

** Associate professor, Faculty of Psychology 1, “Sapienza” University of Rome. Top

*** Psychologist, trainee at the School of Specialisation in Psychoanalytic Psychotherapy, Rome. Top

**** Contract teacher, Faculty of Psychology 1, “Sapienza” University of Rome. Top

1. It is not clear why the Author in her definition did not use the term ‘setting’. What is said about the T.A., in fact, is not different from the idea of setting, which we will discuss below. In psychoanalysis, the only aspect actually shared by the patient and therapist is the setting, in the sense of times, payment and arrangement of the session (couch and armchair). Top

2. On this issue see Codignola (1970). Top

3. Some works on this: Carli (2007a, 2007b, 2008a, 2008b); Carli, Grasso & Paniccia (2007); Carli & Paniccia (2003). Top

4. See Carli & Paniccia (2002). Top

5. We chose WAI-O because,  in its items, there is the recurrence of the words patient as well as therapist. It was important that the associations should be stimulated by the two poles of the relationship, both present in the statements  to be associated on. Top

6. We do not talk about a sample, but a group of psychotherapists. This is due to the fact that the construction of a sample envisages the knowledge of the aspects characterising the universe. We do not feel that the universe of  Italian psychotherapists has yet been defined in its basic variables. Top

7. These are items 5 and 8 for cluster 1; items 2, 3, 4, 7 for cluster 3 and items 10, 11, 12 for cluster 4. Top

8. Items 1 and 9 for cluster 2. Top

9. By the word ambiguity we mean a specific modality, typical of the mind’s unconscious mode of working, of representing the specific objects of reality. On this see Carli, 2007b. Top