Countertransference

Abstract

Countertransference is a type of transference: the therapist's own conscious and unconscious emotional reactions to the patient, filed by MeSH as a psychotherapeutic process and the narrower descriptor directly beneath transference. Freud named it in 1910 as an obstacle to be overcome, but Heimann reversed its meaning at mid-century, arguing that the analyst's emotional response is one of the most important tools of the work. Kernberg then split the concept into a classical reading, limited to the therapist's own unresolved conflicts, and a totalistic one covering the therapist's whole emotional response. A later program led by Gelso and Hayes built rating scales and showed through meta-analysis that reactions relate to worse outcome while skilled management relates to better. Three demonstrations show countertransference as the reciprocal of the template that drives transference itself.

Keywords: countertransference, transference, psychotherapeutic process, countertransference management, therapist self-integration

What Countertransference Is

Countertransference is the redirection onto the patient of feeling, expectation, and behaviour that originate in the therapist. In the MeSH definition it is the conscious or unconscious emotional reaction of the therapist to the patient which may interfere with treatment. It is filed as a psychotherapeutic process, a within-treatment mechanism, and it sits directly beneath transference in the MeSH tree as its one narrower descriptor: transference is the patient's redirection onto the therapist, and countertransference is the therapist's answering redirection back onto the patient (Freud, 1910).

The two concepts are deliberately symmetric, and that symmetry is the reason countertransference is best understood as a special case of transference rather than a separate phenomenon. The same machinery — a relational template laid down in an important past relationship, reactivated by a present person who resembles it — runs in the therapist as it runs in the patient. What makes countertransference distinctive is not its mechanism but its position: it operates in the person charged with helping, so its reactions can either blur the therapist's perception of the patient or, if noticed and understood, sharpen it (Heimann, 1950).

This article follows the concept through a reversal of meaning. It begins as Freud's obstacle, a blind spot in the analyst to be analyzed away. It becomes, in the hands of Heimann, Winnicott, and Racker, an instrument — a form of listening in which the therapist's own reactions carry information about the patient. Kernberg then formalizes the two readings the concept had come to hold at once, and a modern empirical program converts the whole question into something measurable: whether, and under what conditions, the therapist's emotional reactions help or harm the work (Gelso & Hayes, 2007).

From Freud's Obstacle to Heimann's Instrument

Freud introduced the term in 1910, in a short paper on the prospects of psychoanalytic therapy. He described countertransference as arising in the physician as a result of the patient's influence on his unconscious feelings, and he was unambiguous about its status: it was a limitation. No analyst, he wrote, goes further than their own complexes and internal resistances permit, and so every analyst required a personal analysis to recognize and master these reactions. For the first four decades of psychoanalysis this was the settled view — countertransference was a contaminant, evidence of unfinished business in the analyst, and the remedy was more analysis of the analyst (Freud, 1910).

The reversal came in 1950 with Paula Heimann. Where Freud had treated the analyst's feelings as noise, Heimann argued they were signal. The analyst's emotional response to the patient, she proposed, is one of the most important tools of the work: an instrument of research into the patient's unconscious. On this view the analyst's feelings are created by the patient and therefore tell the analyst something true about what the patient is doing, unconsciously, in the room. The task is no longer to eliminate the reaction but to sustain it, attend to it, and read it — to feel the feeling without discharging it into action. Heimann's paper marks the moment countertransference stops being only a fault and becomes, additionally, a mode of perception (Heimann, 1950).

Donald Winnicott sharpened the point a year earlier from a different angle. Writing on hate in the counter-transference, he distinguished the therapist's neurotic reactions, which belong to the therapist's own history, from what he called objective countertransference: feelings, including hatred, that any therapist would justifiably have toward a particular patient given how that patient actually behaves. A mother, Winnicott observed, can love and hate her infant at once, and the analyst working with a very difficult patient must be able to acknowledge a real, warranted hostility rather than deny it — because a denied feeling is far more likely to leak into the treatment than an acknowledged one. The demand this places on the therapist is what Winnicott elsewhere called the holding environment: the capacity to contain the patient's difficult states, and one's own justified negative feelings, without retaliating or withdrawing. Objective countertransference is thus not a failure of the therapist but accurate information about the patient's impact on others (Winnicott, 1949).

Racker, and the Two Directions of Identification

If the therapist's feelings are information, the next question is what kind. Heinrich Racker, working in Buenos Aires, gave the most influential answer in 1957 by analyzing countertransference into two distinct forms of identification. In concordant identification the analyst identifies with the patient's own current experience — the analyst's ego resonates with the patient's ego, feeling sadness with the patient's sadness, fear with the patient's fear. This is the empathic pole, the basis of understanding the patient from the inside (Racker, 1957).

In complementary identification the analyst instead identifies with the patient's internal object — the figure, drawn from the patient's past, toward whom the patient's feelings are directed. A patient who unconsciously casts the therapist in the role of a harsh, rejecting parent may pull the therapist toward actually feeling harsh and rejecting, so that the therapist is drawn into playing the complementary part in the patient's internal drama. Complementary identification is how a patient's relational template reproduces itself in the therapist: the patient does not merely describe the old relationship but re-enacts it, and the therapist's pull to respond in kind is the countertransference signal that it is happening.

The clinical value of the distinction is that it tells the therapist what to do with a feeling. A concordant reaction is a window into the patient's self-experience and can be used, carefully, to deepen empathy. A complementary reaction is a warning: it marks the point at which the therapist is being recruited into the patient's repetition, and acting on it would confirm the patient's expectation rather than interrupt it. Racker's framework connects countertransference to the mechanism of projective identification — the process by which a patient unconsciously induces in the therapist the very state the patient cannot bear to hold — and so gives the therapist's felt experience a definite place in understanding what the patient is enacting (Racker, 1957).

The Classical and Totalistic Definitions

By the mid-1960s the concept was being used in two incompatible ways at once, and Otto Kernberg named the split. On the classical, or narrow, definition — Freud's original — countertransference means only the therapist's unconscious reactions that arise from the therapist's own unresolved conflicts. It is the therapist's transference to the patient, a limited and roughly pathological phenomenon to be analyzed and removed. On the totalistic, or broad, definition — the direction Heimann and Winnicott had opened — countertransference means the therapist's total emotional reaction to the patient, whatever its source, including the reactions the patient's behaviour realistically provokes (Kernberg, 1965).

The difference is not merely terminological, because the two definitions imply opposite clinical attitudes. If countertransference is by definition the therapist's own pathology, then having it is a problem and the response is self-scrutiny and further personal analysis. If countertransference is the therapist's whole response, much of it caused by the patient, then it is an ordinary and expected feature of every treatment and the response is to observe and use it. Kernberg argued for the totalistic view, particularly in work with severely disturbed patients, where the intense feelings such patients provoke are among the clearest data available about their inner world. The totalistic definition has since become the dominant one in both clinical writing and research, though the classical sense survives whenever a therapist's reaction is traced specifically to the therapist's own history (Kernberg, 1965).

The two definitions also mark out what has to be managed. Under the totalistic view the therapist will always have reactions; the therapeutic question is whether those reactions are recognized and contained or enacted unrecognized. The unrecognized case has its own name in the later literature: Jacobs called it a countertransference enactment — a moment in which the therapist, without noticing, acts out the complementary role the patient's transference invites, so that transference and countertransference are jointly performed rather than merely felt. Enactment is the clinical hazard the whole apparatus of management exists to catch, and locating it shifted the field's attention from the therapist's private feeling to the observable behaviour it produces (Jacobs, 1986). That reframing — from whether countertransference exists to how well it is handled — is what made the concept tractable to empirical study, because handling, unlike the unconscious itself, can be rated and correlated with outcome (Gelso & Hayes, 2007).

Measuring and Managing Countertransference

For most of its history countertransference was a concept of the consulting room, described in case reports and defended by clinical authority. Beginning in the 1990s, Charles Gelso and Jeffrey Hayes built a research program to test its claims. Their first requirement was to make the reaction observable. Rather than trying to measure the therapist's unconscious directly, they operationalized countertransference through its visible manifestations — therapist behaviours such as withdrawal, over-involvement, or hostility that trained observers or the therapists themselves could rate — and through the therapist's internal reactions reported after sessions (Gelso & Hayes, 2007).

A parallel effort mapped the reactions empirically. Betan and colleagues surveyed clinicians about their responses to particular patients and factor-analyzed the results, extracting a small set of recurring countertransference patterns — overwhelmed, helpless, positive, special, sexualized, disengaged, parental, and criticized reactions among them — and showing that these patterns track the patient's personality pathology rather than the therapist's idiosyncrasy. The finding supported the totalistic view directly: much of what a therapist feels is a lawful response to the kind of patient in the room, not private noise (Betan et al., 2005). Fauth argued at the same time that the field needed to move from cataloguing reactions toward testing their effects, setting a research agenda around the therapist's internal experience as a measurable moderator of technique (Fauth, 2006).

The decisive results came from meta-analysis. Pooling the accumulated studies, Hayes, Gelso, and colleagues found that countertransference reactions were related to poorer psychotherapy outcomes, and, separately, that stronger countertransference management was related to better outcomes and to weaker countertransference reactions in the first place. Management was operationalized through five therapist factors Gelso and Hayes had proposed — self-insight, self-integration, anxiety management, empathy, and conceptualizing ability — and later captured in the Countertransference Management Scale, a validated instrument for rating how well a therapist handles their reactions (Perez-Rojas et al., 2017). The two meta-analyses together support a single practical conclusion: the reactions themselves are a liability, but the therapist's skill in noticing and containing them is an asset, and the second can offset the first (Hayes et al., 2018).

Management is not one skill but five, and naming them is what makes it teachable rather than a matter of talent. Gelso and Hayes proposed that a therapist handles a reaction well to the degree they possess self-insight, self-integration, anxiety management, empathy, and conceptualizing ability; the Countertransference Management Scale rates each, and reviews of the clinical literature treat the set as the concrete target of supervision and training (Hayes et al., 2011). Table 1 states what each factor names.

Countertransference in Motion

The three demonstrations below make the concept manipulable. The first turns Kernberg's split into a working classifier, showing which of a therapist's reactions count as countertransference under the classical versus the totalistic definition. The second builds Racker's distinction, tracing how a patient's internal object-relationship pulls the therapist toward a concordant or a complementary reaction. The third reconstructs the meta-analytic finding, showing how countertransference reactions and their management combine to predict therapy outcome.

Counted as countertransference under this definition: 5 of 6 reactions. The totalistic definition counts every emotional response to the patient, whether it springs from the therapist's own history or from the patient's behaviour, excluding only purely realistic circumstance.

The definition demonstration makes the classical and totalistic readings concrete. Tagging each of a therapist's reactions by its origin — the therapist's own conflict, the patient's provocation, or realistic circumstance — and toggling between the two definitions recomputes which reactions are counted as countertransference, showing why the same session yields a narrow handful of reactions on the classical view and a much larger set on the totalistic one.

A concordant reaction is a window into the patient's self-experience and can deepen empathy; a complementary reaction marks the point at which the therapist is being drawn into the patient's internal drama, where acting on the feeling would confirm the old pattern rather than interrupt it.

The identification demonstration builds Racker's two directions. Selecting a patient's internal object-relationship and the pole the patient currently occupies traces the reciprocal pull on the therapist: a concordant identification that resonates with the patient's own self-state, and a complementary identification that recruits the therapist into the role of the patient's internalized other.

O = 0.50 + 0.39 × 0.30 − 0.16 × 0.70 × (1 − 0.30) = 0.539. Holding the reactions fixed, raising management shrinks the harm term toward zero while adding its own benefit — the two-part meta-analytic finding that reactions relate to worse outcome and their management to better.

The management demonstration reconstructs the meta-analytic interaction. Setting the magnitude of a therapist's countertransference reactions and their skill in managing them traces the predicted outcome index, reproducing the two-part finding: reactions subtract from outcome, management adds to it, and good management shrinks the harm the reactions would otherwise do.

Worked Example

Take the management interaction the third demonstration draws, and use the effect sizes from the meta-analyses. Model the magnitude of a therapist's countertransference reactions as R and their management ability as M, each on a 0-to-1 scale. Let the harm done by countertransference be the reaction magnitude that survives management, R × (1 − M), scaled by the meta-analytic reaction-outcome coefficient of 0.16; let the benefit of management be M scaled by the management-outcome coefficient of 0.39. The predicted outcome index is O = 0.50 + 0.39 × M − 0.16 × R × (1 − M).

Consider a therapist with strong reactions and weak management: R = 0.70, M = 0.30. The management benefit is 0.39 × 0.30 = 0.117. The surviving harm is 0.16 × 0.70 × (1 − 0.30) = 0.16 × 0.70 × 0.70 = 0.0784. The outcome index is 0.50 + 0.117 − 0.0784 = 0.539.

Now hold the reactions fixed and improve the management: R = 0.70, M = 0.80. The benefit rises to 0.39 × 0.80 = 0.312, and the surviving harm falls to 0.16 × 0.70 × (1 − 0.80) = 0.16 × 0.70 × 0.20 = 0.0224. The outcome index rises to 0.50 + 0.312 − 0.0224 = 0.790. Improving management from 0.30 to 0.80, with the reactions themselves unchanged, raises the predicted outcome from 0.539 to 0.790 and cuts the harm term more than threefold, from 0.0784 to 0.0224. The arithmetic makes precise the meta-analytic message: a therapist cannot always prevent the reactions, but managing them is where the leverage on outcome lies (Hayes et al., 2018).

Discussion

The history of countertransference is a case study in how a concept can invert without changing its referent. Freud, Heimann, and Kernberg were all describing the same thing — the feelings a patient stirs in a therapist — but they valued it oppositely, and the disagreement was ultimately about mechanism. Freud's obstacle and Heimann's instrument are reconciled the moment countertransference is seen as ordinary relational cognition operating in the therapist: the therapist, like anyone, reuses stored templates of past relationships in perceiving a present person, and those reactions are simultaneously a distortion (when they come from the therapist's own history) and a source of information (when they are induced by the patient). Which they are depends on their origin, and telling the two apart is the therapist's central task (Racker, 1957).

The empirical program resolves the old argument in a way neither side anticipated. The classical tradition was right that countertransference can harm — the meta-analytic association between reactions and poorer outcome is real. The totalistic tradition was right that it is universal and usable — management, not absence, is what distinguishes good outcomes. The reactions and their handling are separable, and only the handling is reliably under the therapist's control, which is why the five management factors, and not the suppression of feeling, became the practical target (Hayes et al., 2018). This is also why countertransference matters beyond psychoanalysis: any treatment in which a clinician forms a relationship with a patient will generate it, and it has been analyzed within cognitive-behavioural supervision as much as within analytic training (Prasko et al., 2022).

For cognitive psychology the interest of countertransference is that it locates a clinical construct squarely within the machinery of social cognition, and on the harder side of it. Transference research shows a perceiver reusing a relational template to misperceive a stranger; countertransference research shows the same process in the one person professionally obligated to perceive accurately, and asks what skills let them catch and correct the error in real time. The answer — self-insight, self-integration, affect regulation, empathy, and the ability to conceptualize what is happening — is a portrait of metacognitive control applied to one's own reactions, which is why the concept repays study by anyone interested in how people monitor and regulate the templates they bring to other people (Gelso & Hayes, 2007).

Current Directions

The most active empirical strand refines countertransference from a single quantity into a set of distinguishable types with different consequences for the therapeutic relationship. Building on the factor-analytic maps of the reaction, recent work links particular countertransference types to the moment-to-moment fate of the alliance, showing that some reactions accompany ruptures in the working relationship while others track its repair, so that the therapist's felt experience becomes a real-time indicator of where the alliance stands (Tishby & Wiseman, 2022). This connects countertransference to the wider study of the alliance as a partly trait-like, partly state-like process, in which separating the therapist's stable contribution from session-by-session fluctuation has become a methodological priority (Zilcha-Mano, 2017).

A second strand carries the concept out of its analytic home and into other treatment models and into supervision. Countertransference and its management are now framed as trainable competencies rather than analytic mysteries, with explicit protocols for identifying and working with therapist reactions inside cognitive-behavioural supervision, where the vocabulary was once foreign (Prasko et al., 2022). Across both strands the trajectory mirrors the one transference research took before it: from a singular clinical enigma toward a measurable, moderable, and teachable feature of how a therapist's memory for relationships shapes the treatment they provide, with the therapist's disengaged feelings already shown to blunt the benefit patients draw from transference work (Dahl et al., 2017).

Common Misconceptions

Countertransference means the therapist has a problem.

Only on the classical definition, which restricts the term to the therapist's own unresolved conflicts. On the now-dominant totalistic definition, countertransference is the therapist's whole emotional response to the patient, much of it a lawful reaction to the patient's behaviour rather than a sign of therapist pathology (Kernberg, 1965).

Countertransference should be eliminated.

The modern view is that it should be managed, not abolished. Meta-analysis finds that the therapist's ability to manage reactions relates to better outcome, whereas the reactions themselves are inevitable; the target is self-awareness and containment, not the suppression of feeling (Hayes et al., 2018).

Any strong feeling toward a patient is countertransference.

Racker's distinction matters here. A reaction can be a concordant identification that resonates with the patient's own experience, a complementary identification in which the therapist is pulled into the role of the patient's internal object, or a purely realistic response; the three call for different handling (Racker, 1957).

Countertransference is only a psychoanalytic idea.

It is recognized across therapeutic orientations. The therapist's emotional reactions occur in any treatment built on a relationship, and countertransference and its management are now taught within cognitive-behavioural supervision as well as analytic training (Prasko et al., 2022).

Glossary

Classical (narrow) definition.

Freud's original sense of countertransference as only the therapist's unconscious reactions arising from the therapist's own unresolved conflicts, a phenomenon to be analyzed and removed.

Complementary identification.

Racker's term for countertransference in which the analyst identifies with the patient's internal object, being pulled into the role of the figure from the patient's past toward whom the patient's feelings are directed.

Concordant identification.

Racker's term for countertransference in which the analyst identifies with the patient's own current self-state, the empathic resonance that lets the therapist feel the patient's experience from the inside.

Countertransference behavior.

The observable therapist actions — withdrawal, over-involvement, hostility, and the like — through which countertransference is operationalized and rated in empirical research.

Countertransference enactment.

Jacobs's term for a moment in which the therapist, without noticing, acts out the complementary role the patient's transference invites, so that transference and countertransference are jointly performed rather than merely felt.

Countertransference Management Scale.

A validated instrument developed by Perez-Rojas and colleagues for rating how well a therapist recognizes and contains their countertransference reactions.

Countertransference management.

The therapist's skill in recognizing and containing their reactions, captured by five factors — self-insight, self-integration, anxiety management, empathy, and conceptualizing ability — and related meta-analytically to better outcome.

Countertransference.

The therapist's own conscious and unconscious emotional reactions to the patient; the narrower MeSH descriptor filed directly under transference.

Disengaged countertransference.

A pattern of withdrawn, distant therapist feeling toward a patient; empirically, therapists' disengaged feelings reduce the benefit patients draw from transference work.

Holding environment.

Winnicott's term for the therapist's capacity to contain a patient's difficult states, including the therapist's own justified negative feelings, without retaliating or withdrawing.

Objective countertransference.

Winnicott's term for feelings, including hatred, that any therapist would justifiably have toward a particular patient given how that patient behaves, distinct from the therapist's own neurotic reactions.

Projective identification.

The process by which a patient unconsciously induces in the therapist a state the patient cannot bear to hold, a principal route by which complementary countertransference arises.

Self-integration.

One of the five countertransference-management factors: the therapist's possession of a stable, differentiated sense of self that keeps their own reactions distinct from the patient's material.

Subjective (neurotic) countertransference.

Reactions that originate in the therapist's own history and conflicts, corresponding to the classical definition, as opposed to reactions the patient realistically provokes.

Totalistic (broad) definition.

Kernberg's sense of countertransference as the therapist's total emotional reaction to the patient whatever its source, the now-dominant reading that treats the reaction as ordinary and usable.

Transference.

The redirection onto a present person of feelings and expectations first formed in an important past relationship; the MeSH parent term of which countertransference is the narrower child.

Key Researchers

Sigmund Freud. Founder of psychoanalysis (1856-1939) who introduced the term countertransference in 1910, defining it as an obstacle in the analyst arising from the patient's influence on the analyst's unconscious feelings and requiring the analyst's own analysis to master. Wikipedia - Wikidata

Charles J. Gelso. Counseling psychologist at the University of Maryland (1941-2025) who, with Jeffrey Hayes, built the modern empirical program on countertransference and its management and co-developed the five-factor model of management skill. APA memorial award

Jeffrey A. Hayes. Counseling psychologist at Pennsylvania State University who co-led the meta-analyses establishing that countertransference reactions relate to worse therapy outcome while their management relates to better outcome. Faculty page - Google Scholar

Paula Heimann. Psychoanalyst of the British Psychoanalytical Society (1899-1982) who reversed the meaning of countertransference in 1950, reframing the analyst's emotional response as one of the most important tools of the work rather than a fault to be removed. Wikipedia - Wikidata

Otto F. Kernberg. Psychoanalyst at Weill Cornell Medical College (b. 1928) who distinguished the classical from the totalistic definition of countertransference, arguing that the therapist's total emotional reaction is essential data in work with severely disturbed patients. Wikipedia - Wikidata

Heinrich Racker. Psychoanalyst in Buenos Aires (1910-1961) who analyzed countertransference into concordant identification, feeling with the patient's self, and complementary identification, feeling as the patient's internal object. Wikipedia - Wikidata

Donald W. Winnicott. Pediatrician and psychoanalyst of the British Psychoanalytical Society (1896-1971) who introduced objective countertransference, the justified feelings any therapist would have toward a given patient, in his paper on hate in the counter-transference. Wikipedia - Wikidata

Frequently Asked Questions

What is countertransference in psychology?
Countertransference is the therapist's own conscious and unconscious emotional reaction to the patient. In MeSH it is defined as the emotional reaction of the therapist to the patient which may interfere with treatment, and it is filed as a psychotherapeutic process directly beneath transference, its parent term (Freud, 1910).

How is countertransference different from transference?
They are reciprocal. Transference is the patient's redirection onto the therapist of feelings from a past relationship; countertransference is the therapist's answering redirection onto the patient. MeSH files countertransference as the one narrower descriptor directly under transference, reflecting that the same relational mechanism runs in both directions (Heimann, 1950).

Is countertransference a good thing or a bad thing?
It depends on how it is handled. Meta-analysis finds that countertransference reactions relate to worse therapy outcome, but that the therapist's ability to manage those reactions relates to better outcome. The reactions are largely inevitable; the management is the skill that matters (Hayes et al., 2018).

What is the difference between the classical and totalistic definitions?
The classical, narrow definition restricts countertransference to the therapist's own unresolved conflicts. The totalistic, broad definition covers the therapist's whole emotional response to the patient, whatever its source. Kernberg named the split, and the totalistic view is now dominant in both practice and research (Kernberg, 1965).

What are concordant and complementary identification?
They are Racker's two forms of countertransference. In concordant identification the therapist feels with the patient's own self-state, the basis of empathy. In complementary identification the therapist is pulled into the role of the patient's internal object, feeling as the figure from the patient's past toward whom the patient's feelings are directed (Racker, 1957).

Can countertransference be measured?
Yes. Researchers operationalize it through observable therapist behaviours and reported internal reactions, and factor-analytic studies have mapped recurring patterns that track the patient's personality pathology. The therapist's management of countertransference is captured by the validated Countertransference Management Scale (Perez-Rojas et al., 2017).

How does a therapist manage countertransference?
Through five factors identified by Gelso and Hayes: self-insight, self-integration, anxiety management, empathy, and the ability to conceptualize what is happening in the relationship. These skills let a therapist notice and contain a reaction rather than enact it, and they are teachable rather than fixed (Gelso & Hayes, 2007).

Does countertransference occur outside psychoanalysis?
Yes. Any treatment built on a therapeutic relationship generates it, and countertransference and its management are now addressed within cognitive-behavioural supervision as well as analytic training, framed as competencies to be developed rather than analytic mysteries (Prasko et al., 2022).

References