or to the threats of loss of love his parents have used to bind him to them.21 Children who have received a great deal of genuine affection seem to be those who in later life show in highest degree a sense of security.
In addition to this, it seems probable that another reason for Freud’s misperception of the origins of excessive separation anxiety was the delay in his recognition of the close bond of child to mother and the length of time over which it normally persists at high intensity; only if the child’s strong attachment is perceived as normal is its severance or threat of severance recognized as dangerous. It is true that by the time he wrote Inhibitions, Symptoms and Anxiety he was of opinion that a main cause of man’s proneness to neurosis lies in ‘the long period of time during which the young of the human species is in a condition of helplessness and dependence … (which) establishes the earliest situations of danger and creates the need to be loved’ (1926, pp. 154–155). Yet, so far as I know, he never drew from this the natural conclusion that disruptions or threats of disruption of the primary bond are likely to prove a major hazard.
It will thus be seen that the views advanced in this paper differ from Freud’s not so much on the nature of separation anxiety itself but on the conditions which determine its presence in excessive degree. On this issue indeed the two views are the opposite of one another. It is perhaps because of this and because Freud’s hypothesis of spoiling has been built deep into psycho‐analytic theory that there has been so much reluctance in many analysts to accept as valid the evidence which supports the hypothesis here advanced. It is time to return to this.
In my view the best opportunity for uncovering the conditions which lead an individual to become prone to an excessive degree of separation anxiety is either by direct observation of a child undergoing an anxiety‐provoking experience or by a clinical examination in an analytically oriented child guidance clinic, in which treatment is given to both child and parent and a detailed history can be obtained both of main events in the child’s life and of parental attitudes towards him. When we review the reasons why in some children expectant anxiety in regard to separation and loss of love exist in pathogenic degree, observations made in such settings suggest there are four main ones:
1 One determinant is undoubtedly the actual experience of a period of separation. In addition to our own observations (Bowlby, 1951, 1954; Robertson, 1953a), those of Edelston (1943), Prugh et al. (1953), Heinicke (1956) and Schaffer (1958) provide abundant evidence that the child who returns after not too long a period with strangers, whether in hospital or elsewhere, will soon attach himself with great tenacity to his mother and show intense anxiety at any threat of a repetition of the experience. Many cases of older children and adults who respond to separation with unusual anxiety are most readily understood in terms of the persistence of such a psychological state.
2 Another determinant is the excessive use by parents of threats of separation or withdrawal of love as sanctions.
3 Another is the child’s experience of rejection by the mother, especially where her positive feelings are mixed with unconscious hostility.
4 Another is any actual event, such as a parent’s or sibling’s illness or death, for which the child has come to feel responsible and, therefore, guilty and unloved.
There are many papers by analysts which report cases falling under one or a combination of these last three heads (including an early one of my own: Bowlby, 1940) and others by clinical psychologists.22 In a study predominantly concerned with the consequences of actual separation, however, it would be inappropriate to discuss this large and controversial area more fully. Nevertheless it should be noted that these four sources are not necessarily exhaustive: for example, any set of conditions which results in the child feeling guilty and therefore in danger of not being loved will be effective. At the same time, it is my view that only if each of the four sources listed above has been thoroughly explored and excluded is it wise to postulate other factors. Unfortunately such exploration is, I believe, only possible in the case of younger children and when their mothers are also willing to undertake treatment.
Merely to describe these sources of increased separation anxiety, however, is insufficient: we need also to understand the nature of their effects on the emotional development of the child. It is when we come to consider these effects that the interaction of expectant anxiety and hostility, to which attention has already been drawn, is seen to be so crucial. For each of these experiences – separation, threats of separation, actual rejection or expectation of rejection – enormously increases the child’s hostility, whilst his hostility greatly increases his expectation of rejection and loss. Such vicious circles are a commonplace of psycho‐analytic practice. Since it is in emphasizing their frequency and immense clinical importance that Melanie Klein has made her special contribution, this is a convenient point at which to reconsider her ideas.
The clinical observations made by Melanie Klein in the twenties, it will be recalled, were that some children who are attached to their mother in unusual degree are, paradoxically, also possessed of strong unconscious hostility directed towards that very mother. In their play they demonstrate much violence towards mother‐figures and become concerned and anxious lest they may have destroyed or alienated them. Often after an outburst they run from the analytic room, not only for fear of consequences from the analyst, but also, it seems, to assure themselves that their mother is still alive and loving. These observations are now amply confirmed and demonstrate without doubt that the presence of unconscious hostile impulses directed towards a loved object greatly increases anxiety. This is readily intelligible. As Freud pointed out, we would not expect loss of love or castration ‘if we did not entertain certain feelings and intentions within us. Thus such instinctual impulses are determinants of external dangers and so become dangerous in themselves’ (1926, p. 145). The presence of hostile impulses directed to a parent, especially when unconscious, inevitably increases expectant anxiety. In so far as there is concern for the object’s safety, it is depressive in character; in so far as there is fear of losing his or her love, it is persecutory. The role of such depressive and persecutory anxieties, springing from unconscious hostility, in persons suffering from an increased level of expectant anxiety in regard to being separated or unloved cannot be overemphasized; and this remains so whether or not we accept Melanie Klein’s particular hypothesis in regard to their origin.
But just as unconscious hostility directed towards the loved object increases expectant anxiety, so does expectant anxiety, especially in regard to whether or not one is loved, increase hostility. It is of both great theoretical and great practical importance to determine, if we can, how these vicious circles begin. Does increased anxiety precede increased hostility, is it the other way round, or do they spring from a common source? Jones ([1929] 1948) recognizes the great difficulty of unravelling the sequence when looking backwards from data provided by the patient in analysis; and I believe this holds for young children as well as for older patients. Indeed it is at this point that I believe Melanie Klein’s method has led her to one‐sided conclusions.
Logically it is clearly possible for excess anxiety to precede excess hostility in some cases, for the sequence to be reversed in others, and for them to spring from a single source and so be coincidental in yet a third group. Such possibilities, however, are not allowed for by Melanie Klein’s formulation. It is to be noted that she attaches no importance to instinctual tensions as such and does not subscribe to the view, advanced by Freud and again here, as well as by many other writers, that primary anxiety is the result of such tension. Instead, her basic tenet is that increased anxiety is always both preceded and caused by increased sadism: that it may sometimes be independent of, sometimes itself provoke, and often spring from the same source as the increased sadism is not conceded.
In my view both an excess of separation anxiety and an excess of hostility are very commonly provoked by the same experience. Further and more important is that, because the hostility is directed towards the loved
