In his article on countertransference hate, Winnicott discusses the therapist's reactions to enraging patients, particularly psychotic patients. Winnicott warns that the force and manifestations of countertransference hate in the therapeutic situation must be appreciated. Not being aware of hate as a countertransference response can cause the therapist to act out in ways that can be destructive to the patient. Winnicott discusses how to use this countertransference experience therapeutically with patients. To illustrate potential uses of hate, he describes his interactions with a frustrating young boy who was a runaway during the bombings of London during World War II.
In his case description, Winnicott depicts the boy as the "most loveable and maddening of children" (p. 72). While the child was staying at the evacuation hostel, Winnicott attempted to treat the boy, but with time "his symptoms won," and the boy ran away from the hostel. He later turned up at a police station near Winnicott's home, and Winnicott and his family took in the boy. After a period of giving the boy "complete freedom" (p. 72) during which Winnicott bailed him out of police stations, the boy began to act out more directly at home. Winnicott notes that dealing with his behavior was "a whole-time job" (p. 72).
In reference to the countertransference he experienced with this boy, Winnicott notes, "The important thing for the purpose of this paper is the way in which the evolution of the boy's personality engendered hate in me" (p. 73). He goes on to confess to the reader that though he did not hit the boy, he wanted to. However, "at crises I would take him by bodily strength, without anger or blame, and put him outside the front door, whatever the weather or time of day or night. There was a special bell he could ring, and he knew that if he rang it he would be readmitted and no word said about the past" (p. 73). As he placed the boy outside the door, Winnicott would tell the boy that he hated him. Winnicott goes on to say the disclosure was not only important for the patient's progress, but also to make the situation with the boy tolerable. He also confesses that without these actions he would have been in danger of losing his temper and "every now and again murdering him" (p. 73).
Winnicott subsequently draws parallels between the countertransference hate of the psychotic patient and the hate that a mother can have for her children. An elaborate list of an infant's intrusive qualities is provided to the reader to help in understanding a mother's potential hate. The task, then, for the mother, is to tolerate this hate "without doing anything about it" (p. 74). He warns that "if, for fear of what she may do, she cannot hate appropriately when hurt by the child, she must fall back on masochism" (p. 74). The mother's hate is necessary, according to Winnicott, for the child to tolerate his own hate. This would not be possible in a "sentimental environment" (p. 74). In a parallel manner for the therapist, "the psychotic patient cannot be expected to tolerate his own hate of the analyst unless the analyst can hate him" (p. 74). Winnicott noted that in his interactions with the young boy, the child seemed to appreciate the therapist's interpretation of his running away--that in part it was to flee his own internal world and in part to preserve his mother from his assaults. The boy feared that he might be too powerful.

In Winnicott's discussion of countertransference, he lays out a broad view of this concept, which includes not only the unconscious, repressed reactions of the analyst, but also the "objective" reactions to the real personality of the patient. He proposes that psychotic patients and patients like the delinquent boy evoke justified hate in the therapist, a reaction that is not merely a neurotic response of the analyst. In other words, he is stating that it is normal to hate such patients.
While it may be the case that certain patients are objectively more enraging than others, some authors have been suspicious of Winnicott's pronouncements concerning countertransference hate and his subsequent treatment methods (Blum, 1997; Etchegoyen, 1991). In particular, Winnicott has been directly criticized about his treatment of the boy in this case. Blum (1997) states that Winnicott's case description "includes denial of rage, a pact of silence, and acting out by the analyst. His incomplete exposition suggests collusive, abusive acting out … Hate may be 'justified,' but not the analyst's acting out of countertransference hate" (p. 371). He then raises the question: did Winnicott's countertransference incite the child's hate?
Despite these concerns, most authors hold that the experience of countertransference hate is a normal and even necessary aspect of working with certain patients (Epstein, 1977; Gabbard, 1991; Greenson, 1974; Searles, 1965; Searles, 1986, Bird, 1972; Davies, 2004; Slochower, 1992; Spotnitz, 1976). Yet, the depiction by Blum highlights the confusion and finger-pointing that can develop during the treatment of such patients. This dynamic speaks to the role of projective identification as a relational defense used by the enraging patient (Kernberg, 1984; Gabbard, 1991). With projective identification, conflict-laden aspects of the patient's self are projected and experienced by the therapist (Ogden, 1979). Having the therapist receive such self-states helps relieve the patient of annihilating self-criticism and anxiety that are too much for the patient to manage alone. Thus, the therapist is provoked into being the "bad one" instead of the patient. "It is you, not me," says the patient's unconscious.
If the dynamics of projective identification lead to hate and resentment in the therapist, the question of what to do about it remains. After all, this defense is not only maladaptive for the patient, but also is often unbearable for the therapist. Approaches to the management of the enraging patient have highlighted the use of interpretation, containment, and judicious displays of affect. Interpreting patients' hostile projections has been described as critical in advancing the patient to more mature psychological defenses (Kernberg, 1984). The timing and aggressiveness of interpretation, however, have been the source of debate. Several authors, including Winnicott (1968), warn of the dangers of interpretations in hostile patients and argue for a more gradual use of this technique (Gabbard et al., 1994; Gabbard, 2000; Billow, 1999). Containment by the therapist, on the other hand, is generally accepted as necessary for the reintegration of projections and for the further psychic development of enraging patients (Bion, 1962); Epstein, 1979; Searles, 1986; Gabbard, 1991). During the time of intense patient hostility, the therapist must be careful not to act out on the projections, but manage the clinical sessions despite the intensity of the patient's hateful, projective attacks (Davies, 2004). This can be a difficult balance to hold. As Gabbard (1991) points out, "the analyst must walk a fine line between blasting the patient with his own hatred and denying its very existence" (p. 632). The cautious use of open aggression by the therapist may be necessary to maintain this containment, and this has been supported by authors other than Winnicott (Epstein, 1977; Gabbard, 1991; Hayward & Taylor, 1956; Heimann, 1955; Searles, 1965). Not to show aggression could be detrimental to the patient. Epstein (1977) argues that for the therapist to remain overly benign and seemingly unaffected by the patient's hostility might be seen by the patient as inauthentic and foolish. Furthermore, being too tolerant and compassionate risks inciting envy about the therapist's "good nature" and thereby intensifying destructive impulses. In regard to a hate-inducing patient, Epstein notes "the more benign the treatment she receives in response to her destructive behavior, the worse she would have to feel about herself if she were to believe in the goodness of such treatment" (page 453).
So in light of this brief discussion, how do we understand Winnicott picking up his patient, removing him from the premises, and telling him that he hates him? Is this a judicious use of open aggression while Winnicott functions as a container for the boy's projected, destructive impulses? Or is Winnicott acting out in response to his countertransference in a manner that continues projective identification dynamics? As an introduction to the following case material, I would argue that these questions reference a potential distinction between limit setting and projective identification enactments. Limit setting, as I would define it for this discussion, is the nonambivalent use of aggression to establish tolerable boundaries that enable basic relational functioning and the safeguarding of affection. In reference to the above discussion, it is the appropriate use of aggression for the purpose of containment. With projective identification, aggression is used in a highly ambivalent manner with porous relational boundaries and projection of responsibility. Aggression is not contained in these circumstances but is forcefully uncontained. Limits collapse into a brawl or affective withdrawal. For such patients, this regressive dynamic can create an experience in which their own hostility has far too much power to affect others. A core aspect of the treatment of these patients is the establishment of solid limits that diminish their power and enraging capacity.

T. was a 9-year-old boy who was referred by his pediatrician secondary to problems with physical aggression and out-of-control behavior. At the time of treatment initiation with me, T. was intermittently getting into fights with his mother. This would typically occur when his mother would interrupt a preferred activity or would simply say "no" to the boy. At times, the boy would viciously curse at her and refuse to cooperate with her requests. His mother would not infrequently "lose it," screaming counterattacks at the boy and occasionally physically striking him. She was very angry with her husband, T.'s father, viewing him as siding with the boy and not supporting her authority. He, on the other hand, saw her as being too hard on the child, and he understood the boy's disruptiveness as a result of her own hostility towards T. When the father would attempt to set limits, the boy would rebel intensely and accuse the father of being weak and caving into the mother's demands. These encounters would at times involve the father physically holding T. for several minutes until he calmed. T.'s father noted that he was uneasy with his son's volatility and would minimize demands on him for fear of "setting him off." In addition to fights with his parents, T. was intermittently explosive with teachers at school and occasionally with peers.
T.'s father presented as a reasonable, kind, and supportive man. He had a gentle and caring way about him. I often thought to myself that he would make an excellent therapist. His mother, on the other hand, was critical and sarcastic in her style. I would find myself worried about making mistakes for which I would be reprimanded by her. Initial meetings with his parents were difficult. They recognized that they were split about how to handle T., and they stated that this was not going to change. During parent meetings, I could feel the mother's impatience and frustration with me when progress with T. was unclear. His father, on the other hand, would superficially appreciate my input but frequently fall asleep during appointments.
During individual sessions, T. was angry and difficult to engage. Often when I would greet him in the waiting room, he would be sitting on his mother's lap making bullying demands on her. I would see her hedging about whether to give in, and I could see her resentment rise. He was, at times, reluctant to separate and enter my office, but on almost all occasions he would eventually come in. Once within my office, he was often quiet and angry looking but he would engage with encouragement. During sessions T. could be outright mean, frequently making demeaning comments aimed at me. When participating in sport-related play in my office, he would be much too physically aggressive, and at times he was obviously intending to harm me. I allowed T. to express verbally what he wished to say about me, but I would give him verbal directives to stop when he became too physically aggressive. He repeatedly tested the rules of the office by asking to bring various items home and becoming angry and bullying when told "no." I found myself often worried about his reactions to "no" and felt as if I was "walking on egg shells". I would wonder if my rules were too harsh or strict. In T.'s imaginative play, largely war related, there were often themes of sadistic soldiers who would torture and hurt others. During one session, a soldier, after destroying a whole army, killed himself. At another point, T. made an effigy of his mother out of clay and tortured it repeatedly with a trance-like grin on his face. We began his psychotherapy sessions at once per week, then after six months, increased to twice per week.
My initial formulation of the case depicted the boy's difficulties as a hypervigilant reaction to his mother's abusive behavior. His rage was a justified protest to this treatment. The unfairness of her relational engagement with him continuously provoked him. I saw the origin of her behavior as secondary to her unresolved anger towards her abusive father. T.'s father, on the other hand, I saw as a "good-hearted person" who was unable to protect T. from the unreasonable hostility of his wife. It seemed to me that he thought the situation might worsen if he left the role of the appeaser and protected his son more directly. T.'s father noted that his own father was slow to anger, stating "we could have sawed his arm off and he would not have become angry." During parent guidance meetings, I would try to point out to his mother that her hostile and rejecting attitude towards T. might incite him. I suggested that she lessen her role as a "limit-setter" and let her husband step in with these matters. Most important, I tried to have her envision T. as something other than a young sociopath. This was a hard, if not impossible, sell. I would advise T's father to discuss his disagreements with his wife in a direct manner, when T. was not present. In addition, when the mother would set a limit, he should back her up in the moment and not hedge/undermine as this might provoke T. His father would agree in the office but do little to implement such advice. Both parents were referred to individual treatment. Neither consistently complied with the recommendation.
Most of my initial interpretive statements towards T. highlighted his anger at his mother for being too hard on him and his anger at his father for not supporting him against his mother. I tried to empathize with him that this home situation was very painful at times and his behavioral problems were understandable. During a session near the Christmas holiday, I asked T. if he was worried about being disappointed with the gifts he would receive. He responded by saying "you would worry too if you did some of the stuff I did." I stumbled to lessen his self-criticism by stating that he was too hard on himself and that his difficulties were very much related to problems with his parents. There was little affective response from T. to this comment. In later therapy sessions, I would point out to him that somehow he always was cast as the "bad guy" in his family. This type of comment more than others would temporarily settle him.
With time, T. became more and more provocative in individual sessions. After one year of treatment, he became physically out of control in my office, throwing furniture, requiring me to physically restrain him. When his father eventually returned to my office, he shook his head at T, a look of shame and disappointment crossing his face, then escorted T. from my office. Of note, T.'s father did not look at all angry. With this circumstance, as with others, the father would let T. cool off and then try to speak with him about what had happened. Keeping with my initial formulation, I explained T.'s reaction as ongoing, agitated frustration with his rejecting mother and displaced rage at me for being like his father and not protecting him.
T.'s acting-out behavior escalated at home, with associated suicidal statements that eventually lead to his hospitalization. His mother was critical and disappointed with the "uselessness" of the admission. As per the hospital staff's recommendations, the parents began couple and family therapy with a separate therapist. His parents were inconsistent in following these recommendations. T. continued to have difficulty with disruptive/provocative behavior after discharge, and eventually an outpatient consultation was acquired from a senior psychoanalyst. The consultant recommended child analysis with a new therapist. My role was to shift to that of the psychopharmacologist. I must confess, this was a relief and an obvious sign that the boy and his family had been angering me.

Given the progression of my treatment with T., I sought to understand further his dynamics, and how this contributed to the apparent reenactment within my office. His verbal and ultimately physical attacks towards me were obviously provocative and potentially enraging. I speculated that such behavior was an angry displacement intended as a response to his father's weakness and his mother's hostility. Yet, he seemed to enjoy being provocative and attacking. I felt his behavior towards me was sadistic. Maybe his hostility was not merely all defensive. Assuming that the dynamics that I experienced with T. were a repetition of his primary relationship with his parents, then his parents also may have felt sadistically provoked. This seemed to be the case. Like Winnicott's boy in "Hate in the Countertransference," T. was good at getting others to be angry with him.
On a superficial level, and for the moment ignoring the relational psychodynamics of his parents, T. was able to get others to hate him for a simple reason--he was treating them poorly. T.'s attacks were violating the boundaries and rights of others in an inflammatory manner. It was natural to hate him. So why was T. behaving in such a manner? Again putting a hold on analyzing his parent's deeper dynamics, I would contend that the fundamental environmental reason for T.'s destructive behaviors was his parents horrible inability to set limits.
T.'s mother, despite her focus on T.'s need for limits, would hedge on saying "no." She would frequently give in to his demands and then become resentful, glaring at him as if he were a monster. When she "stuck to her guns," he would intensify his defiance and a brawl would ensue. A typical escalation would proceed as follows:
T. asks for something,
Mother says "no,"
T. calls her accusatory names,
Mother threatens a consequence,
T. continues with attacks,
Mother escalates punishment to unenforceable level,
T. continues verbal attacks,
Mother becomes verbally demeaning and rejecting.
Although his mother may have believed that she was setting a limit, I would hold that her reactions were far from limit setting. She would engage in slug fests with T., who would purposefully attempt to enrage his mother during these altercations. His blaming and calling her an awful mother was too much for her to bear. She had to respond by defending herself and counterattacking. Citing Winnicott, she was "every now and again murdering him" (p. 3).
T.'s father was also poor with setting limits. His timid avoidance of conflict hindered his effectiveness. Most notably, he was unable to summon up anger when a limit was necessary. He would hedge and attempt to diffuse situations in a calm manner. He saw his wife as tyrannical and rejecting, which she often was. But neither could he tell her nor his son, in a firm and committed tone, that their behavior was not tolerable. His limit setting might incite their accusatory responses. Instead, he would cast looks of blame and disgust while affectively withdrawing from his family.
Limit setting was not possible in this family because attempts at it collapsed into projective identification enactments. Exposure to their child's aggression and selfishness (along with other maligned self-states) awakened his parent's self-criticism and intense ambivalence about these qualities in themselves. They would become highly anxious in response to seeing such traits in their son. The mother would react to such self-states in T. by becoming attacking, while simultaneously being unsure as to where to draw the line on such states. When T. counterattacked, accusing her of being mean, cold, abusive, etc., she could not absorb the accusations, but resorted to defensiveness--in essence fighting off her own self-criticism and pitching bad-object status back on to T. To complete the projective identification, T. acted on his mother's projection and identified with such self-states. This was due, in part, to his mother "identifying" him as the source of such despised qualities in the family. It was also due in part to her ambivalence about setting limits on these states, as she was ambivalent about these states in herself. She was not able to offer T. an empathic "no."
T.'s father was also guilty of projective identification enactments. His relationship to his own aggression was obviously anxiety ridden, as evidenced by his overly benevolent and passive stance within the family. When exposed to the rage in his wife and his child--in part his own projected rage--he would act in a fearful manner and affectively withdraw. He was slow to intervene when his son became attacking to his mother, though clearly T. was doing his father's dirty work here. With T. acting as the hostile one in the family, his father was able to "wash his hands" of his own guilt-ridden rage. When eventually he had to set limits on his son's behavior, he had difficulty drawing up enough aggression to withstand his son's defiance and attacks. To set a limit on his son's (and his wife's) power, he needed to use his own power and aggression. As a result, T. continued to be out of control and enraging. Again citing Winnicott, he could not appropriately hate his child, and instead, he fell into a sentimental and masochistic position.
In his sessions with me, T. baited me into a repetition. He would make coercive requests, direct cutting comments, and become physically disruptive. I was uncomfortable around him, as he caused me to feel incompetent and weak. Furthermore, I felt conflicted about how much aggression I should use to contain him. He had me on my heels. I was being pushed into hating him. Unfortunately, my limit setting was too ambivalent. He had me wondering if I was tyrant-like and unfair when I attempted to set a limit, i.e. I was feeling the projective identification dynamics. There was nervousness and self-doubt in my voice during such moments. I needed to feel more comfortable using aggression to directly set limits with T. and to recognize his sadistic attempts to agitate me and project responsibility.

T. began his treatment with a new therapist. Not surprisingly, this treatment was also unsuccessful. The intent for T. to have appointments several times per week never materialized. At this point, his parents set up a return consultation with me. I scheduled several meetings with them. I gave typical behavioral advice about limits. But my psychodynamic formulation added additional depth and direction, primarily by addressing projective identification. I told his father that if anyone was going to be a tyrant in the home, I would want it to be him and not the boy. If his boy hated him momentarily for setting a limit and would tantrum excessively, the father should firmly state his commitment to his limit. His boy was intimidating him, and he was afraid. "Better that he be afraid of you," I stated to him. I assured him that he could handle this type of authority much more than his son (I was taking a leap of faith here). Besides, fearing your child is not good for your child's self-esteem.
I approached things differently regarding T.'s mother. She should try to avoid a defensive brawl with the boy when there was a bad-object tug of war. When he attacked her and accused her of being mean and hateful, she needed to tolerate and ignore the accusations until he settled. I anticipated for her that he would up the ante, increasing provocations as she did this, but she must wait and not counterattack. In parallel, during parent sessions she was very defensive about being blamed. I originally avoided confrontation with her but eventually saw it necessary to be direct. She could be mean and nasty but we could talk about this and work with it. I tried to diminish her self-criticism by making statements such as, "every parent loses it occasionally and this is not all bad." But as Winnicott had emphasized, she must get better at tolerating her own hate. She must take responsibility for her own destructiveness, and then she must set an understanding limit.
During this period of treatment, T. followed his mother in close quarters around the house, while intermittently bouncing a tennis ball against the wall to try to engage her provocatively. He was too big for her to take the ball away physically. He likely would have attacked her. I told her she should tell him to stop it with a firm tone. If he did not, she was to take the ball away figuratively by ignoring it and changing the subject. If she lost her temper, she was to make a genuine but short apology. She was to avoid, as much as possible, becoming hopeless, withdrawn, or resentful. These feeling states of his mother were a sign that the boy and the projected states had again become too powerful.
After this change in formulation and approach, T. continued to test his parents, but his episodes of disruptive behavior and aggression at home greatly decreased. He was less provocative, and there was less defensive brawling. T. eventually was enrolled in a therapeutic day school. T. struggled intermittently with the staff at the school. I was suspicious that under the circumstances his provocations and projective enactments once again became too powerful; this time it was in relation to certain school faculty. Regardless, T. began treatment with another therapist, and I continued as the psychopharmacologist and parent guidance therapist.

In this case, I have highlighted the role of limit-setting failures in relation to projective identification dynamics in the treatment of a provocative child and his parents. It was argued that a fundamental reason for the child's disruptive and enraging behavior was poor limit setting on the part of his parents. Instead, his parents engaged in dramatic struggles with the patient symptomatic of projective identification enactments. The boy's dynamics pressed me into a similar regression in which I hedged around setting limits. This failure contributed to a breakdown in containment of the boy's hostility and ultimately led to a disruption in his treatment. The description has obvious parallels with Winnicott's case presentation in "Hate in the Countertransference," and it is a common treatment scenario for children presenting with disruptive behavior disorders.
As discussed, my response of tiptoeing and avoiding direct confrontation with the patient was a reenactment of dynamics similar to those occurring at home for him. I was not going to be the withholding "bad guy," but rather the benevolent therapist. As Epstein (1977) points out, therapists often aspire to positive feelings towards their patients, such as compassion and warmth, and are uncomfortable being seen as stern, harsh, or cold. However, the avoidance of such affects places the spotlight on the primitive hostility of some patients, potentially inciting a projective battle. In reference to Winnicott, such a stance supports a "sentimental environment" that can make it difficult for these patients to tolerate their own hate. This is likely a common pitfall for therapists working with enraging patients.
In light of this, I would hold that it is important for therapists to consider straightforwardly confronting hostility (verbal, attitudinal, or physical in nature) directed at them. As noted, Winnicott needed to forcefully remove his patient from his premises to protect his own sanity. Not to set limits risks potentially damaging expressions of the therapist's hate. Interpretation, for instance, can serve as a means for aggressive counterattacks and may be a form of unconscious acting out in response to the patient's projections (Epstein, 1979). Interpretations can also be used to tip-toe around the reality of the patient's primary hostility and sadism. This occurred in my treatment with T. By blaming T.'s behavior primarily on the attacks of his mother, I unwittingly bypassed T.'s self-states by saying that someone else was the "bad one." This continued the shame that is associated with this self-state. It ignored the fact that primary aggression may drive such behavior, and that the patient may find sadistic pleasure in it regardless of the other's behavior. It was like saying, "your acts must be rooted in maltreatment by others because only a monster would act in such way on their own." Thus, these interpretations, though compelling and in part accurate, can inadvertently intensify projective dynamics by intensifying the shame and potential self-criticism around unwanted self-states.
Interpretations and clarifications that help the patient feel more comfortable with their aggression may be appropriate. This might involve simple statements that get at shame ridden self-states without attempting to excuse them. "You sometimes like provoking people, don't you", or "you must feel jealous or mean towards others at times", are examples of such declarations that have been helpful in my treatment similar patients. Play therapy settings are full of opportunities to help a disruptive child become more comfortable with his hostility via displacement. It is important that these types of interventions be done with a tone that communicates the therapist's familiarity and comfort with such states. The affective subtext for the therapist might read, "I know what you are trying to do, I know that it is satisfying or compelling for you, I might be overcome with it myself, but let's keep it within reason."
It is important to recognize that those acting on projective identification dynamics may believe that they are setting limits. This was the case with T.'s mother; she saw herself as the enforcer. Her assertions could easily confuse a clinician. Yet, I would hold that limit setting and projective identification enactments look very different. Projective identification enactments, in my experience, present in the following ways:

It is important to consider the role of limiting setting failures in the genesis and ongoing regressions of provocative children with disruptive behavior disorders. Projective identification enactments can undermine effective limit setting with resultant brawling and affective withdrawal. These dynamics may repeat in the therapist's office, leading to further behavioral regressions. The establishment of effective and containing limits by addressing projective identification dynamics can be crucial in the treatment of such children.