SciELO - Scientific Electronic Library Online

vol.27 número4Sindicato Nacional dos Psicólogos: Um caso de mudança“PApi - Pais por Inteiro” Programa de intervenção em grupo para o ajustamento pessoal e a promoção da coparentalidade positiva em pais divorciados índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados




Links relacionados

  • Não possue artigos similaresSimilares em SciELO


Análise Psicológica

versão impressa ISSN 0870-8231

Aná. Psicológica v.27 n.4 Lisboa out. 2009


Resolution of ruptures in therapeuticalliance: Its role on change processes according to a relational approach (*)


Joana Coutinho (**)

Eugenia Ribeiro (**)

Jeremy Safran (***)


This article presents the basic theoretical assumptions of a Relational Approach to Psycho-therapy, particularly in what concerns the interpersonal roots of psychopathology and consequently the way the relational experience therapy provides, may serve to change the client’s dysfunctional interpersonal schema subjacent to symptoms. In the second part of the article we present the clinical implications of the concept of Ruptures in Therapeutic Alliance, seen as a tension or breakdown in the collaborative relationship between therapist and patient. Following Bordin’s conceptualization of the alliance, ruptures may consist of a disagreement about the tasks or the goals of treatment or a strain in the bond. The most important findings that have been collected about the way these interpersonal cycles between therapist and patient can lead to change when efficiently addressed, or to poor outcome or unilateral termination when unresolved, are reviewed. Having already accumulated enough evidence about the importance of the therapeutic alliance, a second generation of alliance researchers is now trying to understand the way the alliance is a mechanism of change. The findings we review suggest that the process of repairing weakened alliances may offer an answer to that question.

Key words: Interpersonal approach, Psychotherapy, Ruptures in therapeutic alliance, Theoretical study.



Este artigo apresenta as premissas básicas de uma abordagem relacional em psicoterapia, nomeadamente no que diz respeito às origens de natureza interpessoal da psicopatologia e consequentemente ao modo como a experiencia relacional que a psicoterapia oferece, pode servir para alterar os esquemas interpessoais disfuncionais do paciente subjacentes aos sintomas. Na 2ª parte do artigo, apresentamos as implicações clínicas do conceito do conceito de Rupturas na Aliança Terapêutica, entendidas como um comprometimento ou quebra na relação colaborativa entre terapeuta e paciente. Seguindo a conceptualização de Aliança de Bordin, as rupturas podem consistir num desacordo ao nível das tarefas ou objectivos do processo ou numa tensão no vínculo. São revistos os resultados mais relevantes que têm sido encontrados sobre o modo como estes ciclos interpessoais entre terapeuta e paciente podem conduzir à mudança quando eficazmente geridos, ou a resultados pobres ou finalizações unilaterais quando não resolvidos. Tendo já acumulado evidência para a importância da aliança terapêutica, uma segunda geração de investigadores na aliança, tem procurado compreender o modo como a aliança é, em si mesma, um mecanismo de mudança. Os resultados que aqui são revistos sugerem que o processo de reparação de alianças enfraquecidas pode oferecer uma resposta a esta questão.

Palavras chave: Abordagem interpessoal, Estudo teórico, Psicoterapia, Rupturas na aliança terapêutica.




When we think in a psychotherapeutic encounter, in a classical individual setting, we immediately come across the idea of two persons sitting together, engaged in the task of observing one of the elements’ internal processes. This relational nature is maybe what best characterizes any psychotherapeutic process.

Within this interpersonal frame of reference, the process of change that occurs in the client, and eventually in the therapist, is better understood through the processes of development and negotiation of the therapeutic alliance.

Our main argument in this article is that despite the theoretical approach adopted in a given therapy, the process of development of the therapeutic alliance, particularly the process of going through moments of impasse and ruptures in the relationship between therapist and client, and resolving them in an efficient way, is the main vehicle of change.

In what follows we will first present the theoretical groundings of our argument, derived from an Interpersonal Perspective and then review the empirical evidence that supports it.


Within an interpersonal approach, any determinant of human behaviour has an interpersonal meaning and is better understood through the principles of human interaction.

According with this perspective almost all human needs and motivations can only be achieved in a social world, and even when we’re alone our internal representations of the others still guide our behaviour. Thus what we call personality must be seen as the social product of the interactions we form and maintain with significant figures in our lives.

This notion that the intrapsychic is structured in a dynamic way from the interpersonal experiences is the central assumption in any interpersonal approach. The centrality of the relational experiences to the self development is a common aspect that is shared by the relational approach and other approaches such as the

British object relations theory, the self psychology, the interpersonal psychoanalysis and the attachment theory. As Ghent (2002) suggested: “The term, relational, was first applied to psycho-analysis by Greenberg and Mitchell back in 1983 when they abstracted the term from Sullivan’s theory of interpersonal relations and Fairbairn’s object relations theory”(p. 12).

There is however important differences between the interpersonal approach and the British object relations theory for example. As Benjamin (1990) so clearly illustrated, in object relations theory the term object itself is a legacy of the classic psychoanalytic intrapsychic theory and Fairburn’s concept of object relations referring to the internalization of the interaction between self and objects had only let us recognize that “where ego is, objects must be”. Benjamin argues that the tendency to collapse other subjects into objects is a problematic aspect in psychoanalysis, one that a relational theory should resolve, by defending that “where objects were, subjects must be”. As the author points out: “the other must be recognized as another subject in order for the self to fully experience his or her subjectivity in the other’s presence” (p. 35). In the same paper Benjamin stressed the differences between a relational approach and self psychology, particularly Kohut’s self psychology. She argues that self psychology has been understanding the parent-child relationship in a one-sided way in that “the self was always the recipient not the giver of empathy”, as if the other would just have the role of stabilizing the self and respond to his needs instead of helping the self to learn how to truly recognize the other and be aware of the outside which is more coherent with a relational approach.

Harry Stack Sullivan (1953), seen as the father of the interpersonal perspective, developed a theory that explains the way psychopathology develops and consequently the way human change may take place.

This Theory of the Interpersonal Introjection (Sullivan, 1953) argues that our self-concept develops through the internalization of the way others communicate with us and about us in the past. In other words, people learn to relate to themselves the same way significant others related to them. Relationships with primary caregivers lead to repertoires of internal models about the self and the world which in turn determine subsequent interpersonal relations. In other words, internal models lead people to engage in interpersonal transactions that confirm them through the dynamics of interpersonal complementarity (Kiesler, 1983). Hence they tend to remain relatively stable throughout the life span (Sullivan, 1953).

Empirical evidence for the notion of stability of internal models comes from longitudinal studies in the attachment theory field. In 2000, Waters, Weinfield, and Hamilton presented three long-term longitudinal studies which assessed infant and adult attachment. The authors found that attachment security was significantly stable in two of the three studies. In all of them the discontinuity in the attachment security was related to salient life events and external circumstances. Another number of studies from the Minnesota parent-child project that has been following families at risk for more than thirty years have been showing that when the contexts keep relatively stable an insecure attachment in infancy is strongly related with behavioural problems in the pre-school and school years and with psychopathology in adolescence (Sroufe, Egeland, Carlson, & Collins, 2005). Hence there’s seem to be contextual variables that determine the degree of stability of internal working models across the life span, suggesting that despite the importance of early experiences, the content of the individual’s internal models may change across the life span. When the early relationships with caregivers and other figures are disturbed, the individual internalizes the unavailability and/or rejection of the other, which manifests itself in the formation of internal schemas of self-destruction and self-judgment.

Despite the use of different terminology, several theoretical orientations, agree that these internal models are directly associated with the affective experience and the maladaptive behavioural patterns underlying psycho-pathological symptoms (Schacht, Binder, & Strupp, 1984).

Psychopathology is seen in terms of recurrent patterns of maladaptative interpersonal behavior, because the internal schemas are acted out in the subsequent interactions the individual participates. When interacting with others, the individual tries to consolidate the image he constructed about himself, thus these confirmatory interactions are complementarity by nature (Kiesler, 1996).

The individual with a psychopathological functioning has a very rigid image about himself and the others, which can only be validated through a restricted set of behaviors from the other. As an example, we may think in someone with a narcissistic personality disorder, whose sense of superiority and grandiosity needs to be continually confirmed by a behavior of submission and admiration by others. These individuals are often perceived by others as someone who coerces them to adopt a particular interactional pattern, which in turn leads to the avoidance of those individuals. Thus there’s usually a vicious circle in which the disturbed individual becomes more and more isolated. This feeling of isolation may be interpreted by the subject as an evidence of his uniqueness and superiority, at a surface level, and at the same time he is confronted with the lack of love and support from others, confirming this way his negative interpersonal schema.

As we mentioned before despite the importance of early experiences and the relative stability of these internal models or interpersonal schemas, due to the dynamics of interpersonal complementarity, their content may change through the lifespan.

The relational experience offered by therapy might constitute one of the contexts in which this change takes place. As other relationships, the one established between therapist and client is the relational stage in which these interpersonal schemas are acted out, therefore the interpersonal transactions between therapist and client may function to perpetuate client’s internal schema, or to disconfirm them through an emotional corrective experience. The concept of emotional corrective experience has its origins in the Franz Alexander, who argued that the fact that the analyst’s reactions are different from that of the patient’s parents is a crucial therapeutic factor for “... it gives the patient an opportunity to face again and again, under more favorable circumstances, those emotional situations which were formerly unbearable and to deal with them in a manner different from the old [Alexander & French, 1946, pp. 66-67]”. About a decade later, Alexander further elaborated the concept, arguing that the analyst should use his knowledge about the patient’s early interpersonal experiences to intentionally assume a different attitude from the parental original one. This new attitude was likely to correct the pathogenic emotional influences of the patient’s early experiences. As Wallerstein (1990) illustrated some authors like Gill saw the concept of emotional corrective experience as proposed by Alexander as not analytic, once the goal of psychoanalysis is an intrapsychic modification in the patient. It is easy to accept that Alexander concept defies Freudian classic psychoanalytic principles of the analyst neutrality. As Gill (cit. by Wallerstein, 1990) noted: “Certainly to meet the patient’s transference behavior with neutrality is to give him a corrective emotional experience without the risks attendant on taking a role opposite to that which he expects” (p. 292).

Relational approaches influenced both by British Fairburn’s object relation theory and American Sullivan’s Interpersonal theory, attribute a central role to the actual patient-analyst relationship to the therapeutic change process. As Fairburn (cit. by Wallerstein, 1990), defended such a relationship with a consistent and trusting figure may function to correct the previous disturbed relationships. It’s easy to recognize that this view resembles Alexander ideas about the emotional corrective experiences though, according to Wallerstein (1990), there are still important differences to note. The new interpersonal relationship therapy offers is also very much valued by relational approaches, but they question the kind of deliberate ability to control the spontaneous contratransference processes advocated by Alexander. Inter-personalists like Hoffman (cit. by Wallerstein, 1990), stress the fact that the therapist is constantly vulnerable to countertransference reactions likely to repeat the patient’s interpersonal patterns.

From what we said so far we may conclude that the therapeutic alliance should not be separated from the technical aspects of therapy. It used to be seen as a pré-condition, that allows the implementation of specific intervention strategies, but within a real Interpersonal Approach, the alliance is by itself an active mechanism of change, for the opportunity it offers to challenge the dysfunctional interpersonal schema.

As Strupp, Butler, and Rosser (1988) pointed out the distinction between specific and non-specific psychotherapeutic factors is erroneous, because differently from a pharmacological treatment in which the biochemistry action may be distinguished from the symbolic meaning of the treatment, psychological interventions can never be disconnected from the relational context in which they’re applied.

Referring to the topic of non-specific factors in therapy, Castonguay (1993) illustrates the distinction between them and common factors, stressing that the alliance constitutes a common factor in therapy, but not a non-specific factor. This is to say that not only the alliance is present in every therapy (dynamic, humanistic or cognitive-behavioural), but it is also a concrete mechanism that helps us understand why people change in therapy. This justifies the importance of therapeutic interventions directly addressing the alliance formation and development.

The concept of therapeutic alliance has its origins in Freud’s early theoretical work on transference (1912). The author pointed out the importance of the positive transference to the success of the analytic process. From Freud’s pioneering work different perspectives on the therapeutic relationship emerged. The origins of the concept of therapeutic alliance are attributed to Elizabeth Zetzel (1956), who saw it as an aspect of the total analysand-analyst relationship based on the capacity and willingness of the patient’s to ally with the analyst and the work of analysis in order to achieve the understanding and cure. She argued that this patient capacity to form a trusting relationship which is essential to the alliance formation, depends on early developmental experiences She was also one of the first authors who pointed out the distinction between the “real” and the transferential aspects of the relationship between therapist and patient.

Influenced by ego analysts’ who focused on the real aspects of the therapeutic relationship, Greenson (1971) developed the notion of the working alliance which is seen as the ability of the patient and therapist to work collaboratively in the treatment goals they pursue. He used the term working alliance to stress the patient’s willingness to actively cooperate in the treatment and his ability to follow the therapist insights and instructions.

Luborsky (1984) also proposed that the therapeutic alliance was one of the curative factors of dynamic therapy. The author defined the strength of the alliance as its capacity to withstand the stresses from internal and external sources without breaking and its degree of persistence and dedication in the therapeutic work for overcoming obstacles in one’s self. Luborsky tried to articulate both the conscious-rational vs. unconscious-transferential aspects in his concept of therapeutic alliance, as well as the facilitative vs. active ingredient dichotomy.

According Safran and Muran (2000), within the relational approaches the concept of the therapeutic alliance it is no longer seen as a reflection of the patient’s transference, instead it is seen as an ongoing negotiation process between two different subjectivities. In other words it is a product of a mutual influence between therapist and patient that occurs at both conscious and unconscious levels. This conception of the alliance has to do the increasing importance of therapist flexibility and spontaneity and of the authentic aspects of the therapeutic relationship within these approaches.

Having already accumulated enough evidence about the importance of the therapeutic alliance as a component of change within psychotherapy (Horvath & Symonds, 1991; Martin, Garske, & Davis, 2000; Safran, Muran, Samstag, & Stevens, 2002), the question clinicians and researchers on this topic try to answer in the present, has to do with the way the alliance can function as a mechanism of change.

Therefore the challenge we have in hands at this moment is the refinement of the hypotheses about how interpersonal transactions, namely those that occur in the context of therapy, produce therapeutic change.

We need to formulate more specific research questions about the relation between alliance and outcome, particularly in what concerns the way that relationship is mediated by the emergence of ruptures and their effective negotiation. Though there’s still a lot of work to be done to clarify this topic, there are already some interesting findings collected by this second generation of alliance researchers.


Because as we said before, the therapeutic relationship is essentially an encounter of two different persons, there are some periods in which the negotiation between these two subjectivities can lead to moments of ruptures and impasse in the relationship. According to Safran and Muran (2000) the notion of intersubjective negotiation proposed by Jessica Benjamim (1990) is a central one when we’re thinking about ruptures in therapeutic alliance. According to the authors, rupture episodes correspond to moments of negotiation between two different subjectivities, thus they can help the patient learn to negotiate the needs of self and the needs of the other in a constructive fashion, without compromising the self or treating the other as an object. This same capacity is referred by Benjamim (1990) as the capacity for intersubjectivity (i.e., the capacity to experience both self and other as subjects) which according to the author is a necessary condition to develop a true capacity for intimacy or authentic relatedness. The assumption is that even in a more directive and structured therapy, the negotiation is present at any given moment in therapy and serves an important human function: the definition of who we are in the relationship with the other.

Benjamin’s notion of intersubjectivity illustrates the process of mutual recognition and regulation in psychotherapy and is inspired by feminist psychoanalytic criticism and Hegel philosophy. According to Hegel (cit. by Safran & Muran, 2000), in order to develop a sense of subjects or the experience of oneself as a self, we need the recognition of the other, but at the same time the other is a danger for us because it threatens our self-sufficiency. Thus the individual tries to control him to assure his sense of independence, however if he controls the other destroying his subjectivity, he can no longer constitute an independent existence necessary to confirm his existence as a subject. Therefore the individual is always caught up in this paradox in which the need for relatedness/proximity colludes with the need for agency/autonomy.

In the therapeutic situation this paradox is also present, and becomes even more evident in moments of ruptures and strains in the alliance, in which both elements are experiencing the tension between the need for recognizing and negating the other as a separate centre of subjectivity. Influenced by Winnicott’s thinking (1969), Benjamin points out that using the other as the object of one’s aggression, can at the end make us experience him as an independent subject, who was able to survive our intent of destruction, and can thus confirm our own subjectivity.

Interestingly, Bordin’s (1979) conceptualization of the Alliance, contemporary to Benjamin’s perspective, as comprising an agreement on therapy tasks and goals and the bond, also stresses the opportunity that therapy offers to clients (and eventually to therapists according to a real interpersonal model) to learn how to negotiate the needs of the self versus the needs of the others. This is a tension human beings have to deal with in every interpersonal situation and many of our clients’ problems come from difficulties in managing this in a satisfactory way.

According to Bordin’s conceptualization we may think of an alliance rupture as consisting in a disagreement about the goals of therapy (e.g., the patient seeks the improvement of his social abilities and the therapist considers the goal should be understand the relation between social anxiety and infantile experiences), about the tasks (e.g., the patient is expecting a more didactic strategy, with the use of role play and modelling exercises and the therapist considers that it is important to adopt experiential strategies as the empty chair technique) and a strain in the bond (e.g., the patient feels the therapist is being critical and not supportive).

All these examples can lead to a deterioration in the relationship between therapist and patient. Moreover all the examples given, illustrate the need to learn how to deal in a constructive way, with the paradox between the need for maintaining relatedness with others and the need for self-definition.

Individuals differ in the way they try to resolve this paradox: there are some clients that privilege the need for relatedness, developing an anxious dependence on others and submitting their own needs and wishes in order to maintain the proximity. With these clients it’s more frequent to detect withdrawal markers of ruptures, in which the patient partially disengages from the therapist his emotions or from some aspect of the therapeutic process (Safran & Muran, 2000).

There are other patients however who privilege the need for self-agency, developing a compulsive self-reliance. They may sacrifice their needs for proximity and care and present themselves in a controlling and dominant way in the relationships. In these cases confrontational rupture markers in which the patient directly expresses anger or dissatisfaction with the therapist or the therapy, are more frequent (Safran & Muran, 2000).

These different tendencies derive from the internal schemas about the self and the world developed in the early relationships with important figures, as mentioned earlier. When these internal schemas are acted out in the therapeutic relationship, the client is “inviting” the therapist to behave in a way that confirms his schema. For example, a very submissive patient who has learned that the expression of anger and other negative feelings can lead to the other’s rejection and abandonment, may present himself in a very deferential way in therapy, coercing the therapist to behave in a more dominant way. As mentioned earlier, this is explained by the principles of interpersonal complementary: submissive behaviour is complementary to dominant behaviour. If therapist responds in a way that confirms the patient’s dysfunctional interpersonal schemas, he participates in maladaptive interpersonal cycles similar to those that occur in the patient’s other relationships (Safran & Segal 1990).

Moments of rupture or impasse suggest thus a critical opportunity to explore and understand the processes that maintain the generalized representations of self-other interactions (Safran & Muran, 2000). As Safran and Muran (2000) argue they are also an entry point to what Greenson (1971) has defined as the central feature of the therapeutic alliance: the collaboration between patient and therapist in the task of observing the patient’s experience. In this perspective the building and repair of the alliance is more than the establishment of a relation to facilitate treatment acceptance. It corresponds to the treatment itself by breaking the interpersonal cycles that maintain the client’s dysfunction.

The building and repair of the alliance can thus be a learning experience in which the client gradually develops a relational schema that represents the other as potentially available and the self as capable of negotiating proximity even in the context of interactive ruptures (Safran & Segal, 1990).

Before review the empirical evidence on the importance of alliance ruptures, we would like to note that the phenomena of alliance ruptures have some communality with other familiar concepts in the alliance literature such as: empathic failures, resistance, transference enactments, therapeutic impasses and negative therapeutic reaction. Because some of these concepts have already been discussed by authors who are interested in their relationship with the alliance, we’ll elaborate very briefly the last one, because it seems to be the less explored when it comes to alliance ruptures.

The concept of negative therapeutic reaction was proposed by Freud. It refers to the patient sense of guilt and masochism based on the prevalence of the death instinct in the economy of psychic life and could manifest itself trough negative reactions to the analysis and the analyst. In this negative reaction to the psychotherapeutic process we may identify some parallels with the concept of alliance ruptures. Because of self-destructive tendencies the patient would experience an unconscious resistance against the improvement that therapy tries to foster. For a person with a very strong sense of guilt and self-destructiveness, improvement would represent a lessening of that self-punishment that he needs. According to Loewald (1972), Freud’s conception of the negative therapeutic reaction is somehow restricted in that the central dimension of the concept was the patient’s resistance against improvement, that is, in Freud’s conception this was not primarily a reaction against the therapist and his efforts.

In a relational approach instincts, like the death instinct manifested in self-destructive tendencies, can no longer be seen as forces seeking discharge enclosed in the psyche of the newborn, but they are to be seen as relational phenomena from the beginning. In other words, the intensity of destructive tendencies in the negative therapeutic reaction would depend, predominantly, on early interactions. Environmental forced have been central to the causation of the negative therapeutic reaction, thus in Loewald’s (1972) words: “the implicit attitude of the analyst as a more benevolent potential superego imago is of importance here”(p. 239). This more interpersonal frame of the concept lead us also to consider the relevance of the therapist counter-transference reaction to the negative therapeutic reaction of the patient, an aspect that according to Loewald was also not sufficiently stressed by Freud’s initial formulation of the concept.


One of the most robust findings in psychotherapy research has to do with the association between the quality of the therapeutic alliance and outcome. In the first meta-analysis of 24 studies Horvath and Symonds (1991) found a correlation of 0.26 and more recently Martin et al. (2000), in an attempt to update the previous metanalysis with several studies that had been conducted more recently, found a correlation of 0.22. The authors stress that although this is a moderate correlation, it seems to be very consistent across different studies and reliable. They also argue that due to the increasing quality of the research on this topic derived from the refinement of the measures, we may rely on these results.

However the relationship between alliance and outcome is not free of controversy mainly due to the limitations of the studies reporting it.

First it is reasonable to think that some methodological aspects may interfere with the relationship found between alliance and outcome. According to Luborsky (1994) some of such factors are: the type of measure that is used (whether it is a self-report questionnaire or an observer judgement); the point of view that is used (patient’s, therapist’s, observer’s); variations in the size of the database used for the alliance measure; the moment in which alliance is measured (whether it is in the initial stages of alliance development or it is measured repeatedly across therapy) and also the length of treatment.

On the other hand the relationship between alliance and outcome is mediated by other variables such as the client and therapist’s personal characteristics and the type of treatment conducted. In what concerns the treatment modality the majority of studies looking at this relationship are still with dynamic therapies, however the alliance seems to be a significant predictor in other therapies as well (Marmar, Gaston, Gallagher, & Thompson, 1989, cit. by Luborsky, 1994).

Concerning the mediating effect of client’s characteristics, research indicates that the quality of early experiences with parents affects clients’ ability to form a working alliance with their therapist (Mallinckrodt, 1991). Also client’s mental health facilitates the formation of the alliance. Goldman (2005) found that the more comfortable a client was with closeness and intimacy, the higher the client rated the working alliance.

On the therapist side, certain characteristics and behaviors (e.g., warmth, flexibility, accurate interpretation) are positively associated with strong alliances (Ackerman & Hilsenroth, 2003), while others (e.g., rigidity, criticalness, inappropriate self-disclosure) interfere negatively with the alliance formation (Ackerman & Hilsenroth, 2001). Also therapists who relate in a hostile manner toward themselves are more likely to act in a hostile way toward their clients. Henry, Strupp, Butler, Schacht, and Binder (1993).

As Barber and colleagues pointed out in 2000, another limitation of most of the studies reporting a relationship between alliance and outcome is the fact that they also do not control the influence of the early improvement in that relationship. Most of them assessed change in outcome without controlling the effect of the early in treatment symptomatic improvement. In order to address that limitation Barber and colleagues in 2000, examined change in outcome from the time alliance was assessed, so that they could take into account the role of previous symptomatic improvement on subsequent symptom change. The authors were able to find for the first time that alliance at sessions 2, 5 and 10 significantly predicted subsequent change in symptoms in dynamic therapy. Their findings suggest that although the alliance early in treatment might be influenced by previous symptomatic change, it is still a significant predictor of subsequent improvement.

Although the findings about the relationship between alliance and outcome do not address the topic of alliance ruptures, we may see them as an indirect sort of evidence of their importance, because if a strong alliance is somehow related to good outcome cases, the process of repairing breakdowns in its quality is supposed to be related to good outcome cases. This proposition is supported by the fact that weakened alliances are associated with dropouts (Samstag, Batchelder, Muran, Safran, & Winston, 1998; Tryon & Kane, 1995).

One of our basic assumptions is that the strength of the alliance varies over the course of treatment, thus decreases in its quality, which according to Safran and Muran (2000) constitute an alliance rupture, are almost inevitable in therapy. Binder and Strupp (1997) in a revision about negative processes in therapy, concluded that the kind of interpersonal process involved in rupture resolution is present in every therapy, independently of the theoretical approach.

Despite the constant presence of moments of impasse or rupture in the alliance in therapy, it is not always easy, for even experienced therapists, to identify them. One of the evidences come from Rennie’s qualitative study (1994) which used the grounded theory to analyse tape assisted recalls of fourteen patients gathered immediately following an hour of therapy. The author found that patients not always reveal their feelings of discomfort or dissatisfaction, presenting themselves in a deferential way in the session. They hide their negative reactions in an attempt to protect the therapist and maintaining the relationship, which suggests that it is very important for therapists to remain attentive to shifts in the alliance, even when they are subtle, and address them in a way that allows the client to explore his concerns without anxiety.

Regan and Hill (1992) results go in the same direction. They asked twenty four patients and respective therapist to report on thoughts and feelings that they were unable to express in treatment, using the Things Left Unsaid Inventory and the Session Evaluation Questionnaire. They then asked the therapists to guess what patients had left unsaid. They found that even experienced therapists were able to identify only 17% of the covert processes of their patients, that is to say, feelings and cognitions they had felt but were not able to express.

Two years later Rhodes, Hill, Thompson, and Elliot (1994) asked nineteen therapists and therapists-in-training to recall misunderstanding events from their own treatment and made a qualitative analysis of the events. Client satisfaction was measured by Client Satisfaction Questionnaire and the addressed vs. unaddressed misunderstanding events was measured by Retrospective Misunderstanding Event Questionnaire. They found that in all the cases, the misunderstanding was associated to one of the following situations: the therapist had done something the client didn’t wanted or needed (therapist gives unwanted advice) or the therapists failed to do something the patient wanted or needed (therapist fails to remember important details). In a resolved misunder-standing event, the patient was able to assert negative feelings and therapist remained flexible and accepting, recognizing his responsibility for the event or changing his behaviour. In contrast, in non-resolved events, patients concealed from their therapists their negative emotions and therapists remained unaware of what was happening until the patient quit therapy.

Therapists’ unawareness of patient’s negative reactions can be detrimental to outcome because therapists cannot explore and deal with client’s reactions they are not aware of. Though even if none of the elements is aware of each other’s covert processes they still interfere with treatment.

On the other hand there are some studies suggesting that therapists’ awareness of their client’s negative reactions is not always beneficial to treatment (Fuller & Hill, 1985; Martin, Martin, & Slemon, 1987). As Safran, Muran, Samstag, and Stevens (2001) argue, we may interpret the results of these studies hypothesizing that therapists become more rigid in their adherence to a specific treatment model instead of addressing the strain in the alliance they just detected, in a flexible and open way, Another explanation the authors point out has to do with the therapists expression of their own negative feelings as a way to cope with their clients dissatisfaction.

This “retaliation” may compromise the alliance and the agenda of the session, and at the same time it may confirm the patient dysfunctional interpersonal schemas of hostility for example. Any interpersonal schema is formed within a relational scenario and contains information of the form: “if I do X others will do Y” (e.g., “if I’m angry others will retaliate”), that’s why an hostile client who goes through this cycle of hostility-counter-hostility in therapy is collecting more evidence that being aggressive is the only way to be in the world.

In a study of change in cognitive therapy, Castonguay, Goldfried, Wiser, Raue, and Hayes (1996) clarified the hypothesis that therapists may become more rigid when they are aware of their client’s negative reactions. In thirty cases of brief cognitive therapy they correlated the outcome measure (Beck Depression Inventory), the Working Alliance Inventory, the Experiencing Scale and the Coding System of Therapist Feedback. They found something unexpected: therapist’s focus on the impact of distorted cognitions of depressive affect was negatively linked with outcome. Conducting a more intensive qualitative analysis of those poor outcome cases, they realized that therapists, when confronted with a rupture, adhered in an even more rigid fashion to the cognitive model, becoming more and more focused on challenging distorted cognitions.

A similar process of therapist’s rigid adherence to the model might had happen in another study of Piper, Azim, Joyce, and McCallum (1991), this time about psycho-dynamic therapy. Sixty-four dyads composed the sample and the treatment consisted of 20 sessions of short term therapy. Therapist intervention Rating System was used to categorize interventions and a comprehensive set of outcome measures was provided by patients, therapists and independent assessors. The authors found that an increased proportion of transference interpretations was negatively associated with both the quality of the alliance and outcome. A subsequent qualitative analysis suggested that therapists may have used transference interpretations to deal with an impasse in the alliance, but the way that strategy was used increased the vicious cycle both therapist and patient were involved. Though these results didn’t consider the adequacy of the interpretation, nor the type of patient or phase of therapy as intermediate variables, they seem to suggest that an inflexible adherence to any specific technique as a way of avoid the exploration of the here and now of the relationship, is counter-productive.

This idea is supported by studies in which the therapist is able to be flexible and open to the exploration of the immediate relational context of the session. Foreman and Marmar in 1985, in a small sample study correlated the California Therapeutic Alliance Scale with patient, therapist and independent ratings of outcome and compared to a list of therapist actions. They found that interpretations focused on client’s defenses against feelings about the therapist or the relationship between both, improved the alliance and were related to cases with good outcome. By contrast interpretations that didn’t address directly the alliance impairment were not helpful.

One year later Lansford (1986) correlated measures of initial alliance, alliance weakness and repair with observer ratings of outcome. The author was able to find an important result: the higher levels of patient alliance ratings were preceded by episodes of rupture and repair in which both elements were able to talk about the interaction and the level of successful resolution of these episodes was related with good outcome. And again more transference allusions were present in poor outcome cases. All the studies mentioned above are more qualitative in nature and tried to detect the emergence of alliance ruptures at a molecular or microscopic level.

However there is another set of studies which address the possible benefits of alliance rupture resolution processes at a more global or macroscopic level, analysing the pattern of development of therapeutic alliance over the course of treatment.

Drawing on theoretical and research literature and using clinical examples Gelso and Carter in a paper of 1994, examined the idea consistent with Mann’s theory (1973) that there are different stages in the process of alliance development. Those stages are: the initial phase characterized by patient’s optimism and positive expectations; an intermediate stage in which the patient questions the value of therapy and its usefulness and finally when this ambivalence is successful dealt with, the patient experiences positive reactions, this time more reality based.

Golden and Robbins (1990) found through the analysis of two successful cases, that patient’s alliance ratings increased, dropped and increased again during the course of the therapy. The authors used the Vanderbilt Psychotherapy Process Scales and the Working Alliance Inventory to determine patterns of alliance development.

Using a quantitative methodology, studies by Patton, Kivlighan, and Multon (1997) and Kivlighan and Shaughnessy (2000), collected empirical support to the hypothesis that a quadratic high-low-high pattern of alliance development was related to better outcome. In the first study Patton et al videotaped sixteen patients and six therapists over two semesters and using hierarchical linear model analysis found that a quadratic pattern of alliance development was present and related to improved outcome. In the second study by Kivlighan and Shaughnessy (2000), the authors used cluster analysis instead of hierarchical linear model, to determine patterns of alliance development which were then correlated with the Inventory of Interpersonal Problems and the Battery of Interpersonal Capabilities. Again the high-low-high quadratic pattern was found to have the greatest association with treatment outcome.

In an attempt to replicate the results of the study mentioned above, Stiles et al. (2004) measured alliance fluctuations in different types of therapy for depression, using data from the Second Sheffield Psychotherapy Project. The Alliance was measured by the Agnew Relationship Measure and outcome was measured by the Beck Depression Inventory and Brief Symptom Inventory. The authors couldn’t find the same U pattern identified by Kivlighan and Shaughnessy (2000) four years before, and none of the four patterns they found was differentially associated with good outcome. However, further analysis lead to the identification of a subset of patients, who went through rupture-repair sequences. These clients with brief V shaped deflections were those who presented better outcomes.

In a more recent study, Strauss et al. (2006) found, in a sample of 30 patients with obsessive-compulsive and avoidant personality disorder receiving cognitive therapy, that the sequences rupture-resolution were significantly related with symptom relieve, both in depressive and personality symptoms, respectively assessed by the Beck Depression Inventory and the Wisconsin Personality Disorders Inventory. These gains were registered even after controlling the effect of the number of sessions and the early in-treatment improvement. The alliance was measured by the California Psychotherapy Alliance Scale.

We may conclude that the investigation of alliance ruptures episodes seems to be a promising research topic for clinicians and academics who believe that the therapeutic alliance is more than a non-specific factor in therapy. We believe that in the future the efforts to replicate with larger samples the findings about the effect size of the alliance on the outcomes should be replaced by the effort to clarify the processes by which the alliance, namely the negotiation of ruptures, plays its role.

The process of alliance development in which ruptures may emerge, and its interaction with the patient’s change process is a multidimensional and very complex one. Thus in order to improve their knowledge of it researchers need to address specific questions such as the way patient’s and therapist’s characteristics interact with the process of alliance formation; the role that the patient’s internal representation of therapeutic relationship play in the change process; the way in which the mutual regulation between therapist and patient in rupture episodes leads to change. We also believe that these research questions might require a shift from larger N quantitative methods to single case designs and different qualitative analysis methods. This might also require a shift from a more molar level of analysis to a molecular one focused on the micro analysis of moment to moment shifts in the interactive process of the therapeutic dyad. As Ackerman and Hilsenroth (2003) suggest: “… it is likely that the most promising strategy for future research may be to examine the interpersonal exchanges between the patient and therapist that impact alliance development. Investigating these in-session interactions may deepen our understanding of the nature of alliance development and the specific variables impacting it” (p. 29).



Ackerman, S. J., & Hilsenroth, M. J. (2001). A review of therapist characteristics and techniques negatively impacting the therapeutic alliance. Psychotherapy Practice and Research, 38(2), 171-185.

        [ Links ]

Ackerman, S. J., & Hilsenroth, M. J. (2003). A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clinical Psychology Review, 23, 1-33.

Alexander, F., & French, T. M. (1946). Psychoanalytic Therapy. New York: Ronald Press

Barber, J. P., Conolly, M. B., Crits-Christoph, P., Gladis, L., & Siqueland, L. (2000). Alliance predicts patients’ outcome beyond in-treatment change in symptoms. Journal of Consulting and Clinical Psychology, 68, 1027-1032.

Benjamin, J. (1990). An outline of intersubjectivity: The development of recognition. Psychoanalytic Psychology, 7, 33-46.

Binder J. L., & Strupp, H. H. (1997). Negative process: A recurrently discovered and underestimated facet of therapeutic process and outcome in the individual psychotherapy of adults. Clinical Psychology Science and Practice, 4, 121-139.

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

Castongway, L. G. (1993). “Common factors” and “nonspecific variables”: Clarification of the two concepts and recommendations for research. Journal of Psychotherapy Integration, 3, 267-286.

Castonguay, L. G., Goldfried, M. R., Wiser, S., Raue, P. J., & Hayes, A. M. (1996). Predicting the effect of cognitive therapy for depression: A study of unique and common factores. Journal of Consulting and Clinical Psychology, 64, 497-504.

Foreman, S. A., & Marmar, C. R. (1985). Therapist actions that address initially poor therapeutic alliances in psychotherapy. American Journal of Psychiatry, 142(8), 922-926.

Freud, S. (1912/1958). The dynamics of transference. In standard edition (vol. 12, pp. 97-108) London: Horgarth Press.

Fuller, F., & Hill, C. E. (1985). Counselor and helpee perceptions of conselor intentions in relation to outcome in a single counselling session. Journal of Counselling Psychology, 32, 329-338.

Gelso, C. J., & Carter, J. A. (1994). Components of the psychotherapy relationship: Their interaction and unfolding during treatment. Journal of Counseling Psychology, 41, 296-306.

Ghent (2002).

Golden, B. R., & Robbins, S. B. (1990). The working alliance within time-limited therapy: A case analysis. Professional Psychology: Research and Practice, 21(6), 476-481.

Greenson, R. (1971). The real relationship between the patient and the psychoanalyst. In M. Kanzer (Ed.), The unconscious today (pp. 213-232). New York: International Universities Press.

Goldman, G. A. (2005). Quality of object relations, security of attachment, and interpersonal style as predictors of the early therapeutic alliance. Unpublished doctoral dissertation, Ohio University, Athens.

Henry, W. P., Strupp, H. H., Butler, S. F., Schacht, T. E., & Binder, S. L. (1993). Effects of training in time-limited dynamic psychotherapy: Changes in therapist behavior. Journal of Consulting and Clinical Psychology, 61, 441-447.

Horvath, A. O., & Symonds, B. D. (1991). Relation between working alliance and outcome in psycho-therapy: A meta-analysis. Journal of Counselling Psychology, 38, 139-149.

Kiesler, D. J. (1983). The 1982 Interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214.

Kiesler, D. J. (1996). Contemporary interpersonal theory and research: Personality, Psychopathology and Psychotherapy. New York Wiley.

Kivlighan, D. M., & Shaughnessy, P. (2000). Patterns of working alliance development: A typology of client’s working alliance ratings. Journal of Counselling Psychology, 47, 362-371.

Lansford, E. (1986). Weakenings and repairs of the working alliance in short-term psychotherapy. Professional Psychology, Research and Practice, 17(4), 364-366.

Loewald, H. W. (1972). Freud’s conception of the negative therapeutic reaction, with comments on instinct theory. Journal of the American Psychoanalytic Association, 20, 235-245.

Luborsky, L. (1984). Principles of psychoanalytic psychotherapy: A manual for supportive-expressive (SE) treatment New York: Basic Books.

Luborsky, L. (1994). Therapeutic alliances as predictors of psychotherapy outcomes: Factors explaining the predictive success. In Horvath A. & Greenberg (Eds.), The working alliance: theory, research and practice (pp. 51-84). New Wiley.

Mallinckrodt, B. (1991). Clients’ representations of childhood emotional bonds with parents, social support, and formation of the working alliance. Journal of Counseling Psychology, 38, 401-409.

Mann, J. (1973). Time-limited psychotherapy. Cambridge, MA: Harvard University Press.

Martin, D. J., Garske J. P., & Davis M. K. (2000). Relation of the therapeutic alliance with outcome and other variables: A Meta-Analytic Review. Journal of Consulting and Clinical Psychology, 68(3), 438-445.

Martin, J., Martin, W., & Slemon, A. (1987). Cognitive mediation in person-centered and rational-emotive therapy. Journal of Counselling Psychology, 34, 251-260.

Patton, M. J., Kivlighan, D. M., Jr., & Multon, K. D. (1997). The Missouri psychoanalytic counselling research project: Relations of changes in counselling process to client outcomes. Journal of Counselling Psychology, 44, 189-208.

Piper, W. E., M., Azim, H. F. A., Joyce, A. S., & McCallum, M. (1991). Transference interpretations, therapeutic alliance and outcome in short term individual psychotherapy. Archives of General Psychiatry, 48, 946-953.

Regan, A. M., & Hill, C. E. (1992). Investigation of what clients and counsellors do not say in brief therapy. Journal of Counselling Psychology, 39, 186-174.

Rennie, D. L. (1994). Client’s deference in psycho-therapy. Journal of Counselling Psychology, 41(4), 427-437.

Rhodes, R. H., Hill, C. E., Thompson, B. J., & Elliot, R. (1994). Client retrospective recall of resolved and unresolved misunderstanding events. Journal of Counselling Psychology, 41(4), 473-483.

Safran, J. D. & Muran, J. C. (2000). Negotiating the therapeutic alliance, A relational treatment guide. The Guilford Press.

Safran, J. D., & Segal Z. (1990). Interpersonal process in cognitive therapy. Aronson London

Safran, J. D., Muran, J. C., Samstag, L. W., & Stevens, C. (2001). Repairing alliance ruptures, Psycho-therapy, 38, 406-412.

Safran, J. D., Muran, J. C., Samstag, L. W., & Stevens, C. (2002). Repairing Alliance Ruptures. In Norcross (Ed.), Psychotherapy relationships that work (pp. 235-254). Oxford University Press.

Samstag, L. W., Batchelder, S., Muran, J. C., Safran, J. D., & Winston, A. (1998). Predicting treatment failure from in session interpersonal variables. Journal of Psychotherapy Practice & Research, 5, 126-143.

Schacht T. E., Binder J. L., & Strupp H. H. (1984). The dynamic focus. In H. H. Strupp & J. L. Binder (Eds.), Psychotherapy in a new key (pp. 65-109). New York: Basic Books.

Sroufe, L. A., Egelend, B., Carlson, E., & Collins W. A. (2005). The development of the person: The Minnesota study of risk and adaptation from birth to adulthood. New York. Guilford Publications.

Stiles, W. B., Glick M. J, Osatuke K., Hardy G. E., Shapiro D. A., Agnew-Davies R., Rees A., & Barkham, M. (2004). Patterns of alliance development and the rupture-repair hypothesis: Are productive relationships U-Shaped or V-Shaped? Journal of Counselling Psychology, 51(1), 81-92.

Strauss, J. L., Hayes, A. M., Johnson, S. L., Newman, C. F., Brown, G. K., Barber, J., Laurenceau, J., & Beck, A. T. (2006). Early alliance, alliance ruptures, and symptom change in a nonrandomized trial of cognitive therapy for avoidant and obsessive-compulsive personality disorders. Journal of Consulting and Clinical Psychology, 74(2), 337-345.

Strupp, H. H., Butler, S. F., & Rosser, C. L. (1988). Training in psychodynamic therapy. Journal of Consulting and Clinical Psychology, 56, 669-689.

Sullivan, H. S. (1953). The Interpersonal Theory of Psychiatry. New York: Norton.

Tryon, G. S., & Kane, A. S. (1995). Client involvement, working alliance and type of therapy termination. Psychotherapy Research, 5, 189-198.

Wallerstein, R. S. (1990). The corrective emotional experience: Is reconsideration due? Psychoanalytic Inquire, 10, 288-324.

Waters, E., Weinfield, N., & Hamilton, C. (2000). The stability of attachment security from infancy to adolescence and early adulthood: General discussion. Child Development, 71(3), 703-706.

Winnicott, D. W. (1969). The use of an object. Inter-national Journal of Psychoanalysis, 50, 711-716.

Zetzel, E. R. (1956). Current concepts of transference. International Journal of Psychoanalysis, 37, 369-375.


(*) Este trabalho integra-se num projecto de doutoramento financiado pela Fundação para a Ciência e Tecnologia com a referência BD27654/2006.

(**) Universidade do Minho, Braga.

(***) New School University, USA.

Creative Commons License Todo o conteúdo deste periódico, exceto onde está identificado, está licenciado sob uma Licença Creative Commons