Psychoanalysis and Psychotherapy

Home What's New

Psychoanalytic Writings

Psychotherapy Service Email Forums and Groups

Process Press Links

Kleinian Studies Ejournal

| Home | Contents | Rationale |Feedback | Interesting Links |


Projective Identification: From Attack to Empathy?

by Simon Clarke
University of the West of England

Abstract: Projective Identification is one of the central concepts in Kleinian psychoanalytic theory. This paper seeks to give a broad overview of the concept, its usage and different interpretations,  and set them in a context which should be useful for students of Psycho-Social and Psychoanalytic Studies. It addresses some of the main arguments and commentaries which surround the concept and attempts to unravel the complexity of the idea, suggesting that it is an invaluable tool in the investigation of the psychodynamics of the social world.

            In previous papers I have outlined what I believe to be a critical sociological  explanation of racism and ethnic hatred (see Clarke,1999, 1999a). It has been in the context of this wider research project that I have introduced and focussed on the Kleinian concept of projective identification. Two strong themes have emerged from this research. First, the complexity of the concept of project identification, and second the scarcity of reviews of this concept, with the exception of the work of Young (1994) within psychoanalytic sociological literature. There is however, a wealth of writing within the psychoanalytic and clinical field, and it from this literature that I will draw on in this paper.

            Several people have remarked to me in the context of academic seminars and conferences that this projective identification `thing’ is all very interesting but for several reasons think it is problematic. First, `It is all very complex’. Second, `It seems to mean many different things and have many interpretations’.  Finally, given these problems, `Is it of any practical use in explaining social phenomena such as racism?’ It is within the context of these remarks that in this paper I seek to provide an overview of the concept and the primary literature in this area. The paper itself, is not meant to be an exhaustive review aimed at psychoanalytic clinicians.  Rather it seeks to clarify some of the main arguments and commentaries in the field which should be helpful for students of psychoanalytic sociology and psychoanalytic studies.

            It should be noted, that often it is difficult, if not impossible to think of projective identification without simultaneously considering the Kleinian notion of position (s), splitting and phantasy which I discuss briefly at the start of this essay (see also Clarke 2001). I then go on to outline the relationship between projection and projective identification before providing a discussion of the work of Klein (1946, 1952, 1993), Bion (1962) and Ogden (1990). I also use the work of  Hinshelwood (1989, 1992), Rosenfeld (1988), Sandler (1987),  and Bott-Spillius (1988, 1985) to provide a critical commentary on the idea of projective identification, and the work of Robert Young (1994) which has been crucial in the formulation of my own writings to provide an array of ideas around the concept. The final section of this paper provides some examples of the use of the concept within sociological analysis. I conclude by arguing, as indeed Young (1994) has done, that projective identification  is one of the most useful psychoanalytic concepts since the discovery of the unconscious.


Between Good and Bad

     Melanie Klein argues in her paper the Emotional Life of the Infant (1952) that from the earliest stages of life, under the threat of phantasies of annihilation emanating from the death instinct, in which the self is felt to be in terrible danger, good and bad are separated. Klein uses the term phantasy to refer to the psychic representation of instinct: `Unconscious phantasies are not the same as day dreams (though they are linked with them) but an activity of the mind that occurs on deep unconscious levels and accompanies every impulse experienced by the infant’ (Klein, 1952, p 251). So, for example, a baby may deal with hunger by imagining the gratification of the breast on one hand, but on another, may feel deprived and persecuted by the breast which denies this satisfaction. These are phantasies – the mental expression of the life and death instincts.

            Klein argues that we develop certain clusters of attitudes, and defences in which we learn to deal with anxiety, terror, love and hate. Describing the paranoid schizoid and depressive positions as ways of coping in the world, Klein postulates that although arising first during the earliest stages of life, we constantly move between positions and this movement under certain conditions continues throughout life. Positions are therefore more than distinct and sequential developmental stages, rather, they are as Hanna Segal (1964) notes `a specific configuration of object relations, anxieties and defences which persist throughout life’ (Segal, 1964, pg xiii).

            The paranoid schizoid position is the earliest form of organisation of the defences in which good and bad are split. Klein argues that there is enough ego at birth for the child to feel anxiety and employ mechanisms of defence. The operation of the death instinct gives rise to fears of annihilation, and because the early ego is particularly weak, it tends to fragment under anxiety. The fear of disintegration emanates from the work of the death instinct and evokes paranoid schizoid defences. The splitting of good and bad is a defence of the primitive ego in which the young child’s fear takes the form of phantasies of persecution. In defence, and in phantasy the world is split into good and bad part objects. The good is introjected and idealised, the bad denigrated, anxiety is projected into something, or someone else – the bad object.

            Projective identification is symptomatic of the paranoid schizoid position in which aggressive expulsions of disowned parts of the bad self occur. We have seen above how anxiety is generated from an internal source in the form of the death instinct, but for Klein, the first form of anxiety from an external force can be found in the experience of birth in which the pain and anxiety suffered are perceived as an attack (Klein, 1952, p62) It is however, for Klein, the child’s experience of feeding which initiates the infant’s first object relation with the mother. The mother is not perceived as a whole object, rather the infant reacts to the experience of something good or something bad, hence the mother’s breast is split into good and bad – satisfying and frustrating.  There is thus a link between intrapsychic forces and environmental factors in the generation of anxiety in the paranoid schizoid position. The differentiation between good and bad enables the infant to experience goodness as a basis for his or her sense of self. In other words, the idealisation and internalisation of the good object forms a basis for trusting and loving:

When the infant feels that he contains good objects, he experiences trust, confidence and security. When he feels that he contains bad objects he experiences persecution and suspicion. The infant’s good and bad relation to internal objects develops concurrently with that to external objects and perpetually influences its course. (Klein, 1952a, pg 59)

            Klein thus describes both the conflict and harmonisation that the child experiences as a result of the interplay between ego, object and experience. There is a constant interaction through the processes of introjection of a good object which functions to re-enforce and defend the weak primitive ego, and projective identification between the internal (object) world and external reality. This form of splitting is characteristic of the paranoid schizoid position in which bad or frightening parts of the self are projected outward and become attached to an external object: the mother’s breast. The problem with this as Klein argues, is the process never entirely fulfils its purpose: `therefore the anxiety of being destroyed from within remains active’ (Klein, 1946, pg 5). Craib (1989) also succinctly notes that the projection of bad into an external object leaves the child in paranoid  fear of external attack - `the fear that bad objects will come back and destroy it from the outside’ (Craib, 1989, pg 146) There is thus a double fear – a fear of being destroyed from both external and internal forces. Anxiety stems from both the social relationship between mother and infant and internal anxiety associated with the death drive. Ogden (1986) stresses the importance of environmental as well as intrapsychic influences on the child through the role of projective identification:

…we can say that Kleinian thinking involves an implicit conception of the importance of environment, although Klein herself may not of fully recognised this implication of the concept of projective identification. Without the mother’s serving as container for the infant’s projective identifications, the infant would be doomed to an autistic or psychotic existence (Ogden, 1986, p37)

            Ogden therefore places far more importance on the intersubjective and environmental experience in the development of the child. Rather than being a prisoner in his or her internal world the child is able to develop in relation to the external world of reality through the mediation of the mother in the role of container (after Bion, 1962) Klein views the paranoid schizoid position as a normal defence against anxiety in early childhood which essentially underpins the primitive ego and allows the transformation to the depressive position. Processes of projective identification, splitting and idealisation are a way of making some order out of chaos for the confused infant. As Klein notes, the introjection (the taking in and internalising) of an idealised good object offers the child protection against persecutory anxiety from which develops a stronger and integrated ego. Thus the tendency to split good and bad lessens as the fear of bad objects diminishes. As the child’s internal and external worlds become less polarised, both good and bad are perceived in whole objects:

His relation to the external world, to people as well as things, grows more differentiated. The range of his gratification’s and interests widens, and his power of expressing his emotions and communicating with people increases. (Klein, 1952, pg 72)

            This for Klein describes the integrated experience of what she calls the depressive position. Conflicts within the self are less likely to be split and pushed into bad objects. There is a recognition of good and bad within the self which allows recognition of this in others who are now perceived as whole objects. This integration experience in the depressive position allows the `conflict between love and hatred to come out in full force’ (Klein, 1952, pg 72). Care develops for others, as does guilt. The individual hates the hating self, seeking to repair, to make reparation for damage done in paranoid phantasy:

When the infant feels that his destructive impulses and phantasies are directed against the complete person of his loved object, guilt arises in full strength and, together with it, the overriding urge to repair, preserve or revive the loved
injured object. (Klein, 1952, pg 74)


            This reparative tendency is for Klein itself a mechanism of defence against anxiety. Klein explains: `My mother is disappearing, she may never return, she is suffering, she is dead. No, this can’t be, for I can revive her’ (Klein, 1952, pg75). There is thus a form of mourning for the lost good object and simultaneously a feeling of guilt for the object destroyed in phantasy. Indeed the infant may be so overwhelmed by depressive anxiety that he or she may deny all love for the object and regress to paranoid schizoid defences. Thus for Klein, in the course of the `normal’ development the child in the depressive position will experience a decrease in the persecutory anxiety associated with the paranoid schizoid position whilst experiencing an increase in depressive anxieties which accompany the desire to make reparation. A variety of anxieties or traumas can cause the infant to move between the two positions. Attaining the depressive position argues Hinshelwood (1989) is a developmental step, but very much an uncertain step. It is therefore a life long task (Hinshelwood, 1989, pg 144

             I have provided this brief outline of the paranoid schizoid and depressive positions as a precursor to the discussion of the main theme of this paper - projective identification. Projective identification is symptomatic of the paranoid schizoid position, of splitting and phantasy. As the infant moves precariously into the depressive position the emphasis shifts from projection to introjection for fear of losing the good loving object. Klein stresses that many of the mechanisms of defence and communication developed in the first years of life continue to be used in adult life and are evoked by certain psychological conditions. As Robert Young (2000) has noted `we are often near, if not over the edge, life is a balance in some sense between love and hate’.


Projecting Onto, Projecting Into

            Projection per se is a process which Freud describes in his monograph Das Unheimlich. I have discussed this paper at length elsewhere (Clarke, 2001) but I feel it is fair to say that this is Freud at his most philosophical – a paper in which Freud turns to aesthetics, to the quality of feelings, not of the sublime or beautiful, but in that wonderful Freudian way, to the uncanny `to what is frightening – to what arouses dread and horror’ (SE, XVII, pg 219) For the sociologist, the best example of the use of the concept of projection in critical theory can be seen in Max Horkheimer and Theodor Adorno’s (1994) Dialectic of Enlightenment,  from it we get the oft quoted phrase `what appears repellently alien, is in fact, all too familiar’

            Projection per se is a relatively straightforward mechanism in which unpalatable impulses, feelings and parts of the self are projected out and onto others. In other words, we project onto the world experiences and qualities that are part of ourselves as if they are part of someone else. Klein's notion of projective identification, however, differs significantly and it is important to differentiate between the terminology at the onset.

             Horkheimer and Adorno's use of projection in explanation of anti-Semitism provides us with a useful insight into the psychodynamic processes that underpin racism and ethnic hatred. But what Horkheimer and Adorno fail to explain is the way in which the recipient of the projection is made to feel, for example, inferior. This is a useful way of differentiating between projection and projective identification. Projection is a relatively straight forward process in which we attribute our own affective state to others. For example, we may feel depressed and view our colleagues in the workplace as being `miserable', or blame others for our mistakes. In contrast, projective identification involves a deep split, a ridding of unpalatable parts of the self into rather than onto, someone else. Julia Segal (1992) elucidates:

            It can be a very powerful means of communication of feelings (used by babies or small children before they can talk, for example). It can also be used as a destructive attack, with nasty or unbearable or `mad' parts of the self evoked in other people in order to destroy their comfort, their peace of mind or their happiness. (Segal, 1992, pg 36).

            Projection per se may not be damaging as the recipient of the paranoid thoughts may be blissfully unaware as such. Projective identification may however involve a forcing of such feelings into the recipient.

            Klein (1952) argues that the processes that underlie projective identification operate in the earliest relation to the breast.  She describes `The `vampire like' sucking, the  scooping out of the breast' (Klein,1952, pg 69), as the infant in phantasy attempts to make his way in to the mothers body. This corresponds to `oral-sadistic' attacks on the breast which are bound up with greed, to introject and empty the mother's body of all that is good and desirable. At the same time, driven by persecutory anxiety, the bad self is projected into the object, the breast:

            split off parts of the ego are also projected on to the mother or, as I would rather call it, into the mother... Much of the hatred against parts of the self is now directed towards the mother. This leads to a particular form of identification which establishes the prototype of an aggressive object-relation. I suggest for these processes the term projective identification. (Klein, 1946, pg 8).

            It is not just bad parts of the self which are split off and projected into others, but also the loving part of the self. Bion’s (1962) concept of `container’ and `contained’ demonstrates the way in which projective identification can be positive and communicative as the child learns from the experience of the interaction between his or her mother. The point for Klein is that a balance can be struck between expulsion of good and bad which is essential for the development of `healthy' object-relations. Various aspects of good and bad objects come closer together. There is a synthesis of good and bad in which objects are seen as whole persons in the depressive position in which there is less of a discrepancy between internal and external worlds as the ego becomes more integrated. Unfortunately for the newcomer to object relations, things start to get more complex the more we read.  For example, Klein claims that forcibly entering an object in phantasy stimulates anxieties which threaten the subject:

            For instance, the impulses to control an object from within it stir up the fear of being controlled and persecuted inside it. By introjecting and re-introjecting the forcibly entered object, the subject‘s feelings of inner persecution are strongly reinforced. (Klein, 1946, pg11).

            For Klein, this situation of introjection and reintrojection of the bad object is at the very heart of paranoia. Examples of such paranoia would be the fear of being imprisoned in another body (or culture) or the fear of another object (or culture) forcing itself inside our `self'. The implication of this is that by using projective identification to expel our `bad bits' into others we live in fear of being invaded by others both internally where other is the object of our phantasy and externally where other is the object of our projections, and thus perpetuate the cycle of persecution. Thus, the notion of projective identification becomes more complex as it is unravelled, and it is because of this complexity that in the following section of this paper I wish to outline some of the main strands of argument and literature connected with this concept.



Projective Identification: From Attack to Empathy?


            Robert Young (1994) suggests that projective identification is `the most fruitful psychoanalytic concept since the discovery of the unconscious' (Young, 1994, pg 120). In a thorough review of the literature, Young points both to the complexity of the nature of projective identification and the numerous interpretations of Klein's material. I have argued that projective identification involves the forcing of a feeling or emotion into some other. In other words the recipient of the projection is affected by the projector of the feelings. Young notes, however, that psychoanalysts are divided on the question `whether or not a real, external Other, who has been affected by projection, is essential to the concept' (Young, 1994, pg 124). Bott-Spillius (1988) takes a wide approach arguing that projective identification is used and directed at both internal and external objects. Hanna Segal (1964) defines projective identification as  projection into an external object: `parts of the self and internal objects are split off and projected into the external object' (Segal, 1964, pg 14). Bion (1962) in Young (1994) talks also of projection into an external object.

            Julia Segal (1992) uses the term to describe a mechanism which is designed to evoke a response in others, `It is a very powerful means of communication of feelings, used by babies or small children before they can talk' (Segal, 1992, pg 36). Joseph (1989 cited in Young (1994) talks of a `subtle nudge' to elicit a response from others. In the clinical setting, the patient may consciously or unconsciously try to induce feelings and thoughts in the analyst to `nudge' the analyst in to a way of acting which is consistent with the patient's projections. This is often very subtle and not consciously intended. The recipient is not always immediately induced to act or behave in a certain way, but importantly made to feel the projections of the projector. Joseph (1989) identifies three types of projective identification within the clinical environment. First, the attack on the analyst's mind: `a kind of total invading' (Joseph, 1989, pg 174). Gianna Williams (1997) likens this form of projective identification to a feeling of being temporarily blinded, as if acid had been thrown in her face (Williams, 1997, pg 928) Second, a partial invasion, a taking over of the analyst's capacities. Finally, putting parts of the self, particularly inferior parts, into the recipient. This we can parallel with Bion's (1962) notion of containment. In the former example it is the patient who becomes identified with analyst's capacities, whilst in the latter, it is the analyst who is identified with the projected bad or inferior parts of the self. Ogden (1986) in some sense supports Bion (1962) in a more communicative than internal account of projective identification,  in the sense that the projection is aimed at an external object and designed to illicit a response both consciously and unconsciously. Bott Spillius (1995) notes that analysts have gradually come to widen their use of Klein's concept of projective identification, noting:

            This is, I believe, one of the few areas in which current Kleinian theory and practice is rather different from that of Klein herself, for we now accept without question the idea that projective identification often, although not always, affects the recipient. (Bott Spillius, 1995, pg 1).


            In other words, phantasy is accompanied with some form of communicative behaviour which has an effect on the external object, the recipient. Thus, Bott Spillius argues, projective identification `communicates something, even when communication is not the projecting individual's primary conscious or even unconscious aim' (Bott Spillius, 1995, pg 1)

            Klein herself uses the term projective identification in several ways as her work evolves but in her paper the Emotional Life of the Infant (1952) she is clear that projective identification plays a vital role in the interaction between internal and external worlds.         


Intrapsychic Projective Identification


            It should also be noted that the process of projective identification is closely bound up in phantasy. Phantasy, as I have defined it in previous sections of this paper, is deeply unconscious at one level, but also an integral part of perception. Projective identification is the vehicle or tool which enables phantasy to become reality by projecting our perception and emotion into others. Projective identification can be viewed as both directed at external and internal objects, but as Young (1994) cautions:

It is important to emphasise that projective identification can occur wholly inside the unconscious of the projecting person and need not be involved at all with behaviour which is unconsciously designed to elicit a response from another person. The Other can dwell exclusively in the inner world of the person who creates the projective identification and supplies the response for his or her phantasy. (Young, 1994, pg 125).


            Joseph Sandler (1987) also offers a critical reading of the usage of the concept of projective identification in Projection, Identification and Projective Identification Sandler traces the historical and progressive development of the concept highlighting three stages of usage. Sandler argues that for convenience, the stages have been delineated but in practice there is considerable overlap between the conceptualisations of projective identification. Sandler's first stage is concurrent with Klein's early writings:

            I speak here of first stage projective identification to emphasise the point that for Mrs. Klein projective identification was a process that occurs in fantasy. Let me put it another way. The real object employed in the process of projective identification is not regarded as being affected - the parts of the self put into the object are put into the phantasy object, the "internal" object, not the external object. (Sandler, 1987, pg 17).

Projective identification therefore is a mechanism that operates within the psychic and internal world of the individual, a process which occurs in phantasy and has little impact on the external object world, a view supported by Robert Young's (1994) writings above, indeed Young is highly critical of Ogden’s (1986) intersubjective approach, arguing that there is too much emphasis on the interpersonal at the expense of the intrapsychic:`In my opinion this impoverishes the concept and does not allow sufficient scope and space for the inner world and internal objects' (Young, 1994, pg 124). Second stage projective identification for Sandler, represents an extension of Melanie Klein's original concept:

            projective identification occurring within the person's phantasy life (reflected in the phantasy distortion of the analyst), can be called first stage projective identification. If either the self or the object represented in such unconscious fantasies is identified with by the analyst to a degree sufficient to contribute to the analyst's countertransference, we have an instance of second stage projective identification. (Sandler, 1987, pg 18).

            In other words, the analyst identifies with the projector's transference phantasy and this is shown in the countertransference of the analyst. Projective identification has simply added another layer to what we regard as transference and countertransference. The transference is not simply a repetition of the past but a result, through projective identification, of the present, of the unconscious relationship between analyst and analysand.

            Third stage projective identification for Sandler, involves a psychodynamic with the external word, with external reality and external objects, and is exemplified in Bion's (1962) and Ogden's (1990) writings:

            The analyst as "container" is, as I see it, the analyst who can tolerate the patient's distress, hostility, and love - indeed, all his phantasies and feelings - and who as a consequence of his reverie can return them to the patient in the form of interpretations which will allow the patient to accept as aspects of himself those parts that he had previously considered dangerous and threatening. (Sandler, 1987, pg 23)

            Projective identification in this form is intersubjective, communicative and exists in both internal and external worlds.

            Although highly critical of Bion’s work Sandler makes a series of points which can only add to our understanding of projective identification. First, Sandler argues that projective identification need not always be accompanied by phantasies of invading and entering. The notion of `into' the object is still commensurable with the idea of projective identification as a self and object representation. Second, for Sandler, projective identification is more than a psychotic mechanism, it is ubiquitous:

            Here I find myself in disagreement with those who tend to see projective identification as a psychotic mechanism only. While in psychosis it may be massive, it is nethertheless ubiquitous. It might also be preferable, in line with this, to speak of pathogenic rather than pathological projective identification. (Sandler, 1987, pg 21).

            Third, when examining second stage projective identification, Sandler argues that it is insufficient to say that the internal phantasy object is put into the analyst. Rather, the patient attempts to actualise the phantasy in a process of transference and countertransference which is based in identification with the patient's phantasy object. There is then a further projective identification into the phantasy analyst object. Finally, and this in the area in which Sandler is most critical, there is the idea of projective identification as communication as exemplified by Bion's (1962) work. Indeed, Sandler argues:

            What I find unacceptable is the notion that this process is one of projective identification, unless the concept is stretched to extreme limits. We would have to say, for example, that the child's cry of distress is "put into" the mother by projective identification, and it seems to me that this represents a caricature of the original concept. (Sandler, 1987, pg 23

            For Sandler, Bion's container model can be separated from the developmental theory in which it is based, as can the concept of projective identification. In other words, for Sandler, in Bion's notion of container and contained there is a lot more going on than simply projective identification. It is at this point that I disagree with Sandler and will in the following sections of this paper demonstrate the usefulness of an intersubjective interpretation of projective identification.

            Although critical of the concept of projective identification, Joseph Sandler's work is important as it reminds us of several problems with this type of concept. First, there are so many different interpretations of projective identification that we have to be careful not to take the interpretation out of context. Second, projective identification is used as both an explanatory concept and a description. If we take the many different meanings of the concept, together with its use as a description then we are open to `pseudoexplanation' (Sandler, 1987, pg 23). In other words, if we are not careful, we tend to see projective identification everywhere in a descriptive form. In order that we may retain the explanatory power of the concept we must lose sight of its specific meaning in the relevant context.

            The problem for the psychoanalytic sociologist is, if we take projective identification to be a wholly intrapsychic phenomena then it is not really of any specific use. We cannot study the complex psychodynamic interactions between individuals, groups and society. As a concept it has only any real use in the consulting room. This position, I believe does no justice to the rich and powerful analytic insights that projective identification can bring to the sociological perspective. In the next section of this paper I will examine what I feel to be the more constructive work of Bion (1962) and Ogden (1986) by looking at projective identification as an intersubjective process, as a form of communication, and indeed, as Bion argues, projective identification as part of the thinking process.


From Internal to External Worlds: Projective Identification as Communication


            Herbert Rosenfeld (1988) distinguishes between two types of projective identification and points toward a third. The first, projective identification as a method of communication is an intensification or distortion of a normal infantile relationship, based in non-verbal communication between mother and infant. The child projects unbearable impulses and anxieties into the mother who is able to alleviate and contain the anxiety the infant feels by modifying or changing her behaviour. In an intensified form of this projection, the patient:

            Projects impulses and parts of himself into the analyst in order that the analyst will feel and understand his experiences and will be able to contain them so they lose their frightening or unbearable quality and become meaningful by the analyst being able to put them into words through interpretation. (Rosenfeld, 1988, pg 117)

            This situation, argues Rosenfeld, is of fundamental importance in that it enables the patient or person to learn to tolerate his or her own impulses and begin to think about experiences which were previously meaningless or frightening - to communicate with others.

            A second form of projective identification identified by Rosenfeld involves the splitting off and projection of anxieties into the analyst with the sole purpose of evacuation. This emptying of frightening and disturbing mental content leads to a denial of psychic reality:

            As this type of patient primarily wants the analyst to condone the evacuation processes and denial of his problems, he often reacts to interpretations with violent resentment, as they are experienced as critical and frightening as the patient believes that unwanted unbearable and meaningless mental content is pushed back into him by the analyst. (Rosenfeld, 1988, pg 118)

            A third form of projective identification aims to control the analyst or object's mind. The patient feels that he or she has forced their way into the analyst. Rosenfeld argues that the projection of `mad' parts of the self often dominates and the projector lives in fear of re-contamination, of counter projective identification in which the analyst will re-invade the patient with his or her own madness. All three of these projective processes may operate simultaneously, and for example someone may use projective identification as communication and control to evoke in the analyst or recipient a `concerned' object.

            Bion (1959,1962) develops Klein's idea of projective identification, introducing the concept of `container' and `contained'; the recipient of the projection acts as a container of feelings, such as love, hate and anxiety.  Bion (1962) explains this model in relation to Klein's work:

            I shall abstract for use as a model the idea of a container into which an object is projected and the object that can be projected into the container: the latter I shall designate by the term contained. (Bion, 1962, pg 90)

            The relationship between container and contained is, as Symington and Symington (1996) note, either integrating or destructive. Depending on the level or degree of aggression in the projective identification the communicative aspect of the projection can range from an attack resulting in destructiveness at one end of the spectrum, to a form of communication that leads to empathy and understanding at the other. Thus, projective identification can be just as constructive as destructive. As Ogden (1990) notes, under `optimal' conditions the recipient of the projection can re-process the feeling evoked and then return it to the projector in a more manageable form, a communicative form. Problems arise when the projection takes the form of a violent expulsion in which the `state of the object is not considered' (Hinshelwood:1994:130).
            Bion provides a theory of emotion which has an environmental or social aspect. Bott-Spillius (1988) argues that the notion of `containment' has lessened the divide between cognition and emotion. As Bleandonu (1994) notes the `container is penetrated and the contained penetrates whenever one emotion replaces another' (Bleandonu, 1994, pg 186). Thus the external object is central in attributing meaning to projected affect, a part of the thinking process. The implication is thus far for Bion that projective identification can be a communicative activity, a way of transmitting meaning by evoking empathy. At another level, projective identification can be a form of violent expulsion. It is, as Bott-Spillius (1988) notes, a way of seeing `thinking' in terms of an emotional experience. In other words, one can learn about one’s self and others through projective identification. Again, projective identification can be `good’ and `positive’. Meltzer (1978) takes this one step further by arguing that Bion's notion of container and contained represents the first `cogent theory of the emotions in the history of psychoanalysis' (Meltzer, 1978, pg 52). We can see projective identification emerge as a positive mechanism, part of learning from experience in which the communicative psychodynamic allows the infant to develop.

            For Bion (1962), therefore, projective identification is part of the thinking process. Originally a procedure for `unburdening the psyche of accretions of stimuli' (Bion, 1962, pg 31) phantasy is projected into the container, and in a reprocessed form projected back into the projector. The point is that bad or intolerable feelings are transformed by the recipient and are made tolerable. Bion calls this process of transformation the `alpha function'. If, as Bott-Spillius (1988) notes, all goes well, then the projector, the infant, eventually introjects this function of transformation and thus develops a means of thinking and tolerance of frustration (Bott-Spillius, 1988, pg 155).

            Ogden (1990) reformulates Bion's ideas and describes a manipulative interpersonal form of projective identification:

            Interpersonal pressure is exerted on the recipient of the projective identification, pressure that is unconsciously designed to coerce the recipient into experiencing himself and behaving in a way that is congruent with the unconscious projective phantasy. (Ogden, 1990, pg 145).

            In Ogden's schema unwanted feelings are dumped into others by inducing in a manipulative way that experience in the recipient, thus altering the behaviour of both projector and recipient. The work of Bion and Ogden is particularly useful for psychoanalytic sociology in that both theorists place an emphasis on the environmental and intersubjective experience of the child. Projective identification become a social psychodynamic which can be as positive as it is manipulative. Ogden’s account of projective identification implicitly involves some form of mediation between the individual, group and society in the creation of social reality.

            The concept of projective identification we have seen is complex with many interpretations. Hinshelwood (1989, 1992) positions all these interpretations on a continuum from a violent prototype of the aggressive relationship to empathy:

            If projective identification varies from expulsion to communication, then at the very furthest point on the benign end of the scale is a form of projective identification underlying empathy, or `putting oneself in another's shoes'... In this case the violence of the primitive forms has been so attenuated that it has been brought under the control of the impulses of love and concern. (Hinshelwood, 1992, pg 133).
            Hinshelwood quite rightly highlights the parallel between this continuum and the movement between the paranoid schizoid and depressive positions. A number of forms of projective identification can be used or employed and these are closely related to the notion of positions. Young (1994) notes that there is `no sharp line to be drawn between normal and pathological, between benign as compared to virulent or malignant projective identification' (Young, 1994, pg 130). Young is not suggesting that we cannot differentiate between normal and pathological projective identification, rather that the various forms and functions of projective identification are often mixed together in a complex way.

            Thus far I have differentiated between projection and projective identification and discussed several forms of this mechanism. These interpretations range from the prototype of a violent aggressive relationship (Klein) to empathy (Bion). I have argued that the most important aspect of projective identification is the nature of the projection, in that it involves projecting into rather than onto the recipient. The implication of this is that the projection has some effect on the object or recipient. Again there has been the site of contentious debate over whether or not there exists a real external Other. Young (1994) notes that the Other can dwell exclusively within the confines of the inner world of  the projector, the projector providing his or her own response to the projection. Segal (1992) argues that projective identification is a powerful means of communication that is designed to evoke a response in others. Joseph (1994) talks of a subtle `nudge', whilst Bion (1962) expands on the communicative aspect of projective identification in his notion of `container' and `contained'. Meltzer (1978) and Bott-Spillius (1988) highlight the importance of Bion's work in terms of thinking about emotion within psychoanalytic theory. Ogden (1990) voices most strongly the interpersonal and communicative aspect of projective identification in which manipulation and coercion are central to the way in which we make others feel.

            It would be unwise and unfruitful to view projective identification as any less than a combination of the notions that I have discussed. As Hinshelwood (1989) notes these interpretations can be placed on a continuum, I would suggest that different forms of projective identification are used in differing psychological situations and this is relational to the paranoid schizoid and depressive positions. Despite this complexity, there are clearly different forms of projective identification which I have outlined in this paper. I have argued that these forms can be usefully categorised as intrapsychic and intersubjective projective identification. Whilst intrapsychic projective identification provides us with a valuable insight into early childhood development, it is the intersubjective account of projective identification that provides a potent tool for the examination of social psychodynamics. I have also noted that projective identification can be both good, that is constructive and communicative, as easily as it can be pathological and destructive. This I believe is implicit in Klein's work. Bion (1962) and Ogden (1990) add depth to Klein's work, expanding on both interpersonal and intrapsychic elements of object relations.


Projective identification, Racism and Sociological Analysis


            Several sociologists have suggested that the use of psychoanalytic ideas may be useful in the explanation of racism and ethnic hatred. Rustin (1991) draws our attention to the work of Jean-Paul Sartre (1976) and Klein (1946), locating racism and communal hatred in phantasy and paranoia. Rustin stresses that psychoanalytic interpretations are not sufficient in themselves but are crucial in the examination of the irrational mental processes that uphold social and political structures. Frosh (1989) attributes racism to the `safety of the psyche’, arguing that outpourings of racial violence are a reflection of the racist’s own fear of disorder directed at others in defence of the self. Indeed for Frosh:

The racist defence, along with the phantasy of `masculine’ order, is part of the hatred of all that modernity brings – of its terrors and disconnections, of its promise and its fertile creativity. Racism, consequently, is not just anti-Semitic or anti-black; it is anti-world, anti desire, anti-modernity itself. (Frosh, 1989, pg 243)


            While reminiscent of a form of Nietzschean nihilism, Frosh’s work starts to lay a foundation for a psychoanalytic interpretation of racist disorder. Elliott (1994) similarly tries to chart the relevance of psychoanalysis, and particularly Kleinian theory, to contemporary cultural criticism. Elliott argues that the Kleinian view of love and hate, of human pain, anger, and despair, is of considerable importance to critical social analysis, since destructive and negative feelings intertwine with socially valorised racial attributions. The racist splits the world in rigid categories – the social and political world are characterised by the splitting of good and bad.

            Robert Young (1994) has identified a `loud silence’ in the psychoanalytic literature on racism. With the exception of the authors already cited and the work of Fanon (1968), Kovel (1970) and Wolfenstein (1981) there have been few attempts to apply psychoanalytic ideas, and in particular Kleinian insights in any systematic way. Young draws our attention to some of the psychological characteristics of racial hatred which include: splitting, violent projective identification and stereotyping.

            In the study of racism we are interested in the form of projective identification that is meant to evoke a response in others, to make them feel in a certain way. An analogy we can draw on is the difference between prejudice and discrimination. If prejudice is about pre-judging in the absence of evidence and discrimination is acting on prejudice, then prejudice on its own is relatively harmless; it is action that hurts. Similarly, projections and projective identifications that are contained and lived out in the inner world have little impact on the external world. In which case they would seem to have little sociological importance. It is the violent expulsion, the attack, the impulse `to suck dry, bite up, and rob' to make others feel in certain ways, which provides the explanatory power in projective identification.    Building on this work, I have argued in previous papers (1999, 1999a, 2000, 2000a) how projective identification can enhance our understanding of the psychodynamics of ethnic hated.

            For example, I have suggested (Clarke, 1999a)  that Frantz Fanon (1968) in Black Skin White Masks gives some powerful examples of the lived experience of the `victims' of racism in an analysis of the psychic consequences of colonialism which resonates with the mechanisms of projective identification. Fanon's work concentrates on the psychology of oppression and on strategies to resist oppression.  Fanon's understanding of the psychology of oppression  is that inferiority is the outcome a double process, both socio-historic and psychological:`If there is an inferiority complex, it is the outcome of a double process: primarily economic; subsequently, the internalization, or better, the epidermalization of this inferiority' (Fanon,1968, p13). When Sartre talks of anti-Semitism as a passion it is not the Jewish person who produces the experience; rather, it is the projected identification of the Jew which produces the experience. Experience is made reducible to the object. Sartre agues that Jews have become poisoned by the stereotype others have of them, they live in fear that their actions will correspond to this stereotype and conduct is "perpetually overdetermined from inside" (Sartre:1976:95). Fanon illustrates this internalization of projection:

     My body was given back to me sprawled out, distorted, recoloured, clad in mourning in that white winterday. The negro is an animal, the negro is bad, the negro is ugly.." (Fanon, 1968, pg 113).

            If we refer to the breaking up bodies, to distortion, as more than a metaphor, then these processes are the outcome of projective identification. The white person makes the black person in the image of their projections, as Fanon notes:

            the white man has woven me out of a thousand details... I was battered down with tom-toms, cannibalism, intellectual deficiency, fetishism, racial defects, slave ships..."(Fanon, 1968, pg 112).

            The black person lives these projections, trapped in an imaginary that white people have constructed; trapped by both economic processes and by powerful projective mechanisms which both create and control the Other.

            Similarly, Paul Hoggett's (1992) study of white peoples resentment of Bangladeshi communities in Tower Hamlets, London, illustrates the projective mechanisms that give rise to ethnic hatred. Social dislocation and a series of stabbing's created a profound anxiety amongst the white working classes in the area leading to tension between the two communities. The cockroach became a focus of paranoia and defensiveness. Despite evidence suggesting that a major cockroach infestation had very plausible structural causes -  the improvement of homes by introducing double glazing and central heating, the white tenants refused to believe this to be the case. This small insect came to represent a complex body of resentment, fear and hatred. Indeed for Hoggett:

The resentment the whites feel toward the Bangladeshi community is made poignant by the fact that the latter community has many characteristics - extended and extensive kinship networks, a respect for tradition and male superiority, a capacity for entrepreneurship and social advancement - which the white working class in the area have lost" (Hoggett, 1992, pg 354).

            In this sense, phantasied elements of the white working class community are projected into the Bangladeshi community. The white working class project their demoralised state into the Bangladeshi community in the form of hostility toward the lifestyle adopted by the community while simultaneously experiencing a loss of their way of life.

       I have also suggested the way in which projective identification is useful in understanding the construction of colonial identity (2000a). Using again, the work of Sartre (1976) and Fanon (1968) I have argued that phantasy provides a vehicle for the construction of identity and otherness, indeed the internalisation of projected phantasy is implicit in Fanon’s work. Black identity, is a false identity for Fanon. It is a construction of colonial phantasies about otherness. The black person is literally battered down by the projective identifications of the white oppressor. Black identity has been so powerfully constructed by white culture that the black person adopts a `white mask’

            We hate the group we have constructed in our own minds; racism is a fear of our projected phantasy. In some sense the use of the concept of projective identification to illustrate the violent expulsion of our own unpalatable thoughts, fears, and phantasies into some other is fairly straightforward, and avoids, or at least does not fully credit the very powerful analytic nature of the concept. We tend to associate racism and hatred with very real and tangible acts of physical violence and spatial boundaries. There is also the question of the more subtle, but no less damaging forms of racism which can seen as acts of psychic violence. This is typically the case with, for example, institutional racism, where projective identification is often subtle, a gentle nudge, almost covert but nether-the-less the recipient is made to feel uncomfortable, anxious and different. These forms of projective identification induce in the recipient a feeling of not belonging to an institution and ultimately a sense of exclusion. In a study of the experience of black and Asian students in higher education (see Clarke, 2000) I have argued that there is evidence of  subtle forms of projective identification in which ethnic minority students are made to feel that they do not belong in the institution by their white peers. There is a complex interaction between the projection of individual anxiety at one level and the macro social environment of the institution as a promoter of projective identification at another level. This is reinforced by stereotypical constructions of `black’ and `difference’ which induce in the student a feeling of not belonging. If we are to address institutional racism then we need to understand the subtle processes and emotional mechanisms which underpin this form of social exclusion. Kleinian theory, and particularly the concept of projective identification can give a valuable insight into these processes, and provide some purchase on the visceral and embodied nature of racism.



            I have written this paper in such a way that it provides an overview and introduction to some of the most important issues (and some of the less well known) surrounding the concept of projective identification. It is aimed at the student of psycho-social and psychoanalytic studies and will also hopefully be of some interest to clinicians and practitioners. The key question would seem though: is this concept of any use to sociologists? My immediate response would be a resounding “yes“. The complexity of the concept should not stand as a barrier to the rich and valuable insights that Kleinian psychoanalytic theory can give us into the equally complex social world that we live in. I have tried to unravel this complexity by illustrating the different forms of projective identification in terms of intrapsychic and intersubjective interpretations of the  concept. I have suggested that it is the intersubjective form of projective identification that is most useful in the field of psychoanalytic sociology without undermining the value of intrapsychic formulation in the clinical setting.

            In previous papers (1999, 2000, 2000a) I have used the notion of projective identification to analyse the work of Fanon (1968), Zizek (1993) and Sartre (1976) on colonialism and ethnic hatred and have drawn attention to these works in the previous section of this paper to give a brief illustration of the usefulness of the concept and to guide the reader to further literature in this area. Although projective identification  has given me a certain purchase on the affective and visceral elements of racism and compliments sociological ideas. There are also some very clear ways in which psychoanalytic insights can help us understand the process of research methodology. For example, it may help us understand the projective dynamics of the interview process and the psychodynamic construction of the research environment by both interviewer and respondent. I conclude and concur with Robert Young when he argues that projective identification is one of the most useful psychoanalytic concepts since the discovery of the unconscious and would urge students of psycho-social and psychoanalytic studies to engage with the complexity of the idea of projective identification so that they might enjoy the rich and valuable insights the concept gives in the analysis of the social world that we live in.


Works Cited:


Bion, W.R. (1959). Experiences in Groups. London: Tavistock.

__________ (1962). Learning From Experience. London: Karnac Books.

Bleandonu, G. (1994). Wilfred Bion: His Life and Works 1897-1979.                                                               London: Free Association Books.

Bott Spillius, E. (1988). Melanie Klein Today: Developments in theory and Practice. Volume 1:                               Mainly Theory. London: Routledge.

_____________ (1995). Introduction to the Topic of Projective Identification. Conference:                            Understanding Projective Identification: Clinical Advances: UCL. 28-29 October.

Clarke, S. and Bird, J. (1999a). Racism, Hatred and Discrimination through the Lens of                                     Projective Identification. Journal for the Psychoanalysis of Culture and Society.                            4 (2) pp. 158-161.

Clarke, S. (1999). Splitting Difference: Psychoanalysis, Hatred and Exclusion. Journal for the                                    Theory of Social Behaviour. 29 (1) pp. 21-35.

_________ (2000). Experiencing Racism in Higher Education. Journal of             Socio-Analysis. 2 (1).                                    pp. 47-63.

_________ (2000a). Psychoanalysis, Psychoexistentialism and Racism. Psychoanalytic                                              Studies. 2 (4). pp. 343-355.

_________ (2001). From Aesthetics to Object Relations: Situating Klein in the Freudian                                      Uncanny. Free Associations. Forthcoming.

_________ (2001a). The Kleinian Position: Phantasy, Splitting and the Language of                                         Psychic Violence. Journal for the Psychoanalysis of Culture and Society. 6 (2).

Craib, I. (1989). Psychoanalysis and Social Theory: The Limits of Sociology.                                                       London: Harvester Wheatsheaf.

Elliott, A. (1994). Psychoanalytic Theory: An Introduction. London: Blackwell

Fanon, F. (1968). Black Skin White Masks. London: MacGibbon & Kee.

Freud, S. (1961). The Uncanny.  The Standard Edition of the Complete Psychological Works of                                 Sigmund Freud vol. XVII (1917-1919). London: Hogarth Press. pp 219-252.

Frosh, S. (1989). Psychoanalysis and Racism. Richards, B. (ed.) Crisis of the Self.           London: Free                           Association Books. pp. 229-244.

Hinshelwood, R.D. (1989). A Dictionary of Kleinean Thought. London: Free Association Books.

________________ (1992). Clinical Klein. London: Free Association Books.

Hogget, P. (1992). A Place for Experience: A Psychoanalytic Perspective on Boundary,                                Identity, and Culture. Environment and Planning D:Society and Space. 10. pp.                          345-356.

Horkheimer, M. and Adorno, T. (1994). Dialectic of Enlightenment.London: Continuum.

Joseph, B. (1989). Projective Identification:  Some Clinical Aspects. Felman, M.             and Bott                             Spillius, E. (eds). Psychic Equilibrium and Psychic Change.London: Tavistock/                             Routledge. pp. 168-180.

Klein, M. (1946). Notes on Some Schizoid Mechanisms. In (1993) Envy and Gratitude and                              Other Works 1946-1963. London: Karnac Books. pp. 1-24.

________ (1952). Some Theoretical Conclusions Regarding the Emotional Life of the Infant. In                         (1993) Envy and Gratitude and Other Works 1946-1963. London: Karnac Books.                         pp. 61-93.

________ (1952a). The Mutual Influences in the Development of Ego and Id. In (1993) Envy                            and Gratitude and Other Works 1946-1963. London: Karnac Books. pp. 57-60.

________ (1993). Envy and Gratitude and Other Works 1946-1963. London: Karnac Books and                              the Institute of Psycho-Analysis.

Kovel, J. (1970). White Racism: A Psychohistory. New York: Pantheon.

Meltzer, D. (1978) The Kleinian Development. London: Karnac Books.

Ogden, T.H. (1986,1990). The Matrix of the Mind: Object Relations and the                                                       Psychoanalytic Dialogue. London: Karnac Books.

Rosenfeld, H. (1988). Contribution to the Psychopathology of Psychotic States: The Importance                   of Projective Identification in the Ego Structure and the Object Relations of the                   Psychotic Patient. Bott Spillius, E. (ed). Melanie Klein Today: Developments in                     Theory and Practice, Volume 1: Mainly Theory. London: Routledge. pp 117-137.

Rustin, M. (1991). The Good Society and the Inner world. London: Verso.

Sandler, J. (1987). The Concept of Projective Identification. Sandler, J. (ed). Projection,                              Identification and Projective Identification. London: Karnac Books. pp. 13-26.

Sartre, J-P. (1976). Anti-Semite and Jew: An Exploration of the Etiology of Hate. New York:                           Schocken Books.

Segal, H. (1964). Introduction to the Work of Melanie Klein.                                                                         London: Heinemann.

Segal, J. (1992). Melanie Klein. London: Sage.

Symington, J. & N. (1996). The Clinical Thinking of Wilfred Bion.             London: Routledge.

Williams, G. (1997). Reflections on Some Dynamics of Eating Disorders:             `No Entry' Defences                                    and Foreign Bodies. International Journal of Psycho-Analysis. 78. pp. 927-941.

Wolfenstein, E. (1981). The Victims of Democracy: Malcom X and the Black                                                         Revolution. California: UC Press.

Young, R.M. (1994). Mental Space. London: Process Press.

___________ (2000) Psychoanalytic Pioneers: Melanie Klein. Unpublished lecture notes.                            London, Tavistock Clinic, 26 January.

Zizek, S. (1993). Tarrying with the Negative.Durham: Duke University Press.


Simon Clarke is Associate Director of the new Centre for Psycho-Social Studies at the University of the West of England and Faculty Research Fellow. His research interests include the interface between psychoanalytic and sociological theory; Kleinian and Post-Kleinian thinking; the social application of psychoanalytic theory and practice and psycho-social research methods. He has published widely on the psychoanalytic understanding of racism and ethnic hatred and is currently completing a new text book which explores Psychoanalytic and Sociological accounts of these phenomena.

Address for Correspondence:

Dr Simon Clarke
Centre for Psycho-Social Studies
Faculty of Economics and Social Science
University of the West of England
Bristol, BS16 1QY, UK




 | Home | What's new | | Psychoanalytic Writings | Psychotherapy Service | Email Forums and Groups | Process Press | Links |