The therapist in this vignette faces a familiar dilemma. He is encountering intense transference anger based on what he feels is a misperception on the part of the patient. Some research (1) demonstrates that patients with borderline personality disorder, in contrast with comparison subjects, attribute negative qualities to neutral faces, which they may regard as threatening and nefarious. However, the choice regarding whether or not to address the patient’s emotional reaction through transference interpretation presents a significant dilemma for the therapist. While useful in the “here and now” of the present moment, a transference interpretation can also be experienced as a form of criticism and may arouse more anger of the kind it aims to address.

Should transference interpretation be central in the work of the psychotherapist treating the patient with borderline personality disorder? This question raises fundamental controversies about the therapeutic action of dynamic psychotherapy. While transference interpretation used to be regarded as the central therapeutic factor in analytic therapy, today there is wide acceptance that there are multiple modes of therapeutic action and the therapeutic relationship may be as important, if not more important, than interpretation of the transference (2). On the other hand, an effective transference interpretation may be crucial to strengthening the therapeutic alliance (3).

The essence of transference is that intrapsychic factors—representations of figures from one’s past and the feelings associated with those figures—shape the patient’s perception and interpretation of experience, leading to stereotyped or rigid responses. Hence Ms. A assumes that her therapist does not care about her when he asks for clarification about store policy. Transference interpretation is classically defined as making something conscious to the patient that was previously unconscious—specifically, that the patient’s attributions of certain qualities to the therapist derive from past figures. In this manner insight is provided about patterns that operate outside the patient’s awareness. Transference interpretation is modified when working with patients who have borderline personality disorder. For example, one may focus on the establishment of a common view of reality in the therapeutic relationship or point out how aspects of the relationship are handled by splitting. Hence even though the patient may be consciously aware of both dimensions of the split, the therapist may still interpret the function of the splitting: “Last week you said you hated me, but today you describe me as the smartest therapist in the world. Is it possible that you still have some of those feelings of hate today but are simply too afraid to express them?”

We suggest that a false dichotomy is made between an exclusive focus on transference interpretation as the mode of therapeutic action and an avoidance of transference interpretation in favor of focusing on strengthening the therapeutic relationship. We assert that it is the juxtaposition of an increasingly well-defined therapeutic alliance with inevitable transference enactments that helps clarify problematic ideas and feelings and provides a here-and-now situation in which the patient and therapist can together negotiate how to counteract them with more adaptive (because realistic) ones.

Our purpose in this article is primarily clinical—we briefly examine the research literature on borderline personality disorder to see what we can draw from these studies that is helpful to the dynamic therapist. Psychotherapists who turn to empirical studies on the subject will find preliminary data but not definitive answers. Høglend et al. (4) conducted a randomized controlled trial of dynamic psychotherapy designed to determine the impact of a moderate level of transference interpretations (one to three per session) in once-weekly therapy of 1 year’s duration. They found that patients with low quality of object relations had somewhat better outcomes with transference interpretation than without, but the sample included so few patients with borderline personality disorder that little can be concluded about this approach in this patient group.

If we examine borderline personality disorder in particular, we find that there are at least six forms of psychotherapy that have demonstrated efficacy in randomized controlled trials: mentalization-based therapy (5), dialectical behavior therapy (6), transference-focused therapy (7), schema-focused therapy (8), supportive psychotherapy (7), and Systems Training for Emotional Predictability and Problem Solving (9). While schema-focused therapy, with its emphasis on self-other beliefs and the differentiation between reality and projection, has much in common with dynamic approaches, we confine our discussion primarily to mentalization-based therapy and transference-focused therapy, the two explicitly psychodynamic therapies. One of these modalities, transference-focused psychotherapy, focuses specifically on transference interpretation. In a head-to-head comparison of transference-focused therapy, dialectical behavior therapy, and supportive psychotherapy, transference-focused therapy did slightly better than the other two in terms of increasing reflective functioning and moving the patients from an insecure attachment classification to a secure one (7, 10).

Both dynamic treatments are efficacious for borderline personality disorder, and both treatments improve mentalizing. However, they approach transference work differently (11). Mentalization-based therapy deemphasizes the provision of insight through transference interpretation because the developers of this therapy are concerned that transference interpretation, especially of anger, may destabilize borderline patients (11). In addition, mentalization-based therapists worry that interpretation of transference carries the risk of conveying that the therapist does all the thinking, thus short-circuiting the patient’s mentalizing process. Mentalization-based therapy focuses instead on the patient’s current mental state and mental functioning, helping the patient develop a greater sense of self-agency and more introspective capacity.

Transference-focused therapy, on the other hand, views unintegrated anger as central to borderline psychopathology, and transference interpretation is designed to help the patient integrate anger as well as its associated object and self representations into whole object representations (or configurations of role relationship models) rather than split-off part representations (or dissociated role-relationship models).

Given the conflicting arguments, how can a psychodynamic therapist understand how so many different approaches appear to work? There are several possible answers. When one examines the treatment manuals of these therapies, an overlap is clear. Both types recognize that state control is important and is gained through clarifying, priming, and using a therapeutic alliance to help the patient increase self-observational capacities for reappraisal of erroneous ideas (12). All therapeutic approaches provide a systematic conceptual framework that allows patients with borderline personality disorder to organize their internal chaos and make sense of it.

A second possibility is that different types of patients with borderline personality disorder respond to different elements of the therapeutic action in the diverse psychotherapies. Some patients may readily use transference interpretation, while others may experience enough state destabilization that they cannot process the ideas offered in a realistic appraisal. These effects in a study using group mean values for outcome can lead to a “washout” effect, as those who benefit cancel out those who do worse (13). Further research is needed to tailor the type of therapy to specific patients. Careful moderator analyses of existing studies might be useful in this regard.

We know that a good enough therapeutic alliance is a key ingredient in outcome. Although variously defined, the core concept involves the patient’s perception that he or she is being helped and is pursuing goals held in common with the therapist (3). Norcross (14) noted from his analysis of numerous approaches to psychotherapy that technique accounts for only 12%–15% of outcomes across all different kinds of therapies. The therapeutic relationship transcends technique and is the best predictor of outcome, but well-timed and judiciously used transference interpretations may help counteract a patient’s negative perception of the relationship, leading to a better therapeutic alliance. In the therapy modalities called nondynamic, when one follows the process and manuals, it is noted that therapists in cognitive and behavioral therapies also act to counteract obstacles presented by the patient that interfere with motivation to cooperate with the therapist’s offer of specific methods for improvement in thought and action.

Based on clinical experience and psychotherapy research findings, most clinicians would agree that the alliance between therapist and patient is a key predictor of outcome in all treatments for borderline personality disorder. Most would also agree that the therapeutic alliance may be difficult to stabilize with these patients. In addition, most would acknowledge that the sheer quantity of transference interpretations may be far less important than the timing of such interventions.

Timing and word choice are vital in determining whether a transference intervention facilitates rational reappraisal or destabilizes information processing in favor of the patient’s experience of raw but confusing affect. In a psychotherapy process study involving audiotapes of long-term dynamic psychotherapy with borderline personality disorder patients in which the impact of the therapist’s intervention on the therapeutic alliance was rated (3, 15), transference interpretation was noted to have a greater impact on the therapeutic alliance—both positive and negative—than other interventions. The investigators concluded that transference interpretation is a high-risk, high-gain intervention for borderline personality disorder patients (3). They noted that transference interpretation is most effective when the road has been paved with a series of empathic, validating, and supportive interventions that create a holding environment for the patient. We would add here that the road continues beyond transference interpretation to other methods of facilitating useful change in enduring beliefs.

Combining the results of empirical studies with clinical knowledge derived from following mini-outcomes of change in the midst of therapies, we can summarize the road paved before transference interpretations as well as the road that leads beyond to further change in core attitudes. We summarize this journey as phases of therapy, and it can be illustrated by charting the course of therapy with Ms. A. We think these successive phases offer a systematic way to organize the therapist’s task, which can often seem murky and mired in chaotic transferences and countertransferences. We recognize that these phases are fluid, rather than strictly linear, and that they overlap to some extent. However, we are presenting them sequentially for the sake of clarity.

The first phase clarifies a pattern that is maladaptive. The therapist challenges the patient’s acceptance of the pattern as “ordinary” to her (which it is) and appropriate (which it is not—there is the challenge). The therapist clarifies intention: the challenge is in the long-range interest of the patient. Although it may feel insulting, it is not a scornful or malevolent stance.

The initial phase encourages the patient to see her distress as growing out of her particular perception of events rather than as an accurate replica of reality. Ms. A might be asked what she imagined the clerk was thinking. This attempt to understand someone else’s mind as harboring different beliefs from those in the patient’s mind is the essence of mentalizing. The therapist identifies the recurrent interpersonal attitude as recognizable to the patient intrapsychically as a state of mind. In the vignette the patient might, for example, name her own state with the therapist as a “chip-on-the-shoulder” state. In this state the self is seen as an aggrieved potential victim while others are regarded as mean-spirited, scornful, and potentially harmful. This state can shift with trigger stimuli into an undermodulated rage that can destroy positive interpersonal relationship possibilities, whether with strangers, intimates, or the therapist. Hence the therapist also works in a more general way to help the patient reflect on thoughts and feelings before acting on them.

This initial phase includes mentalizing, clarifying cognitions, and focusing on the mental operations of attention control that might serve to govern behavior at the point of the shift from embittered to enraged states. The therapist maintains calm and challenges the patient to see how often there is an internal readiness to see others in this particular way when in fact this might not be the case. He might say, for example, “This chip-on-the-shoulder state seems to appear in different situations with considerable frequency—in the store the other day, and even with me when I asked a question about it.”

Once the patient can see this pattern, she may still feel unable to change the attitudes that produce the self-impairing states. The therapist patiently helps her gradually identify an underlying role relationship model within these chip-on-the-shoulder states and within the self-righteous states of undermodulated rage (12, 16). The key here is the recognition of the chip-on-the-shoulder state because, when in this state, Ms. A might be able to retain sufficient reasoning processes to prevent entry into the rage associated with counterattack.

As she becomes familiar with the role relationship model in which she felt ready to be treated malevolently, she could become familiar with the chronic nature of her readiness to feel aggrieved through interpretations that link this model to past situations. These interpretations aim to delink this childhood role relationship model from its application to current interpersonal situations. Again, the therapist patiently clarifies and challenges the patient’s perceptions and reactions to show that some are irrational current appraisals of others.

The first phase is state analysis and control, and the second maintains an alliance while challenging and clarifying underlying fantasies and beliefs about self and other that lead to maladaptive patterns. The third phase is to hold on to clarity about the irrational and continuing feeling of having been aggrieved and to link it to Ms. A’s personal life story. For example, one might offer the following observation: “It sounded to me that when you complained to your parents about their neglect, they would verbally blast you for it. Maybe that led you to reverse roles and engage in verbal counterattacks from a position of strong feelings rather than a position of weakness.” In this way the therapist provides an explanatory statement that leads up to transference work without directly addressing the transference relationship.

This observation is repeatedly contrasted with the reality of the growing therapeutic alliance: in fact, the therapist is not verbally abusing her by challenging her to reappraise an interpersonal situation. Rather, he is clarifying that her fantasy of being abused is a reenactment, not a reasonable response to his supposed neglect, lack of empathy, or abuse. This phase may usefully include well-timed transference interpretations. These are likely to promote rather than destabilize the patient’s sense of security and coherence because of the context of a solid therapeutic alliance that has been forged from the first two phases.

The road needs to continue from here, however, into a fourth phase. Here the insights gained, the agreements negotiated, and the underlying sense of alliance are used to form a rationale for counteracting the chip-on-the-shoulder state of mind and the attitudes that organized it. This phase is like the techniques advocated in virtually every form of psychotherapy: using rational “cognitions” to counteract habitual, almost knee-jerk, assumptions and fantasies that are reflective of dysfunctional beliefs and internal object relations patterns or person schemas (17).

The patient will still tend to repeat the maladaptive pattern. In the fifth phase, the mentalization capacities learned are used to consider the possibilities of alternative perspectives in the minds of others. Change is encouraged by considering other possibilities that might explain the behavior of people such as the clerk. The therapist emphasizes that the patient needs to be ready for the emergence of the habitual sense of grievance and to maintain alertness so she can detect pre-rage signals from herself and the trigger behaviors she expects from others.

This intention to counteract habitual irritability and misinterpretation of interpersonal signals includes a practiced readiness to be more tolerant, patient, and reflective. This attitude may need to be faked initially as a way of relating with ordinary social courtesy and of avoiding any sense of humiliation. Specifically, this shift may mean that Ms. A will have to work in therapy to develop an understanding that other people might be surly or irritable for internal reasons and not because of inherent malevolence toward her in particular. The willingness to assume a stance of practicing this strategy is instrumental in deescalating from the state of being aggrieved. It also may assist the patient in being at least benignly indifferent to the surly behavior of another person.

We suggest that many effective interventions in the dynamic therapy of patients with borderline personality disorder share the end goal of changing intrapsychic core attitudes about self and other for the purpose of altering maladaptive interpersonal relationship patterns. These techniques may involve a combination of transference interpretation, clarification and building of a therapeutic alliance, and encouragement of mentalization. All of these maneuvers involve the clarification of specific belief dysfunction and the strengthening of alternative concepts of self and other.

The road we suggest starts with state control, building of observational and thinking skills, and reinforcement of a therapeutic alliance. Along the road, transference interpretations may provide a here-and-now crucible of mutual observation in which to negotiate attitudinal change and further strengthen the alliance. Insight into developmental origins of the transference reactions and their current situational co-construction is not enough. The road goes beyond that to developing new counterattitudes, intentionality, and practiced preparations for new and better encounters with others.