contributes
to the infant's feeling that she will never return'.
Bowlby - Separation
It is held
____________________
1 A version of this review was published in the Journal of Child Psychology and Psychiatry,
Vol. I, 1961. Only few changes have been made; papers published since 1960 are not as systematically considered as those published earlier.
-375-
in many variants according to how the traumatic situation to. be avoided is conceived.
Principal variants are: (a) the traumatic situation is an economic disturbance that is caused when there develops an excessive accumulation of stimulation arising from unsatisfied bodily needs ( Freud 1926a); (b) it is the imminence of a total and permanent extinction of the capacity for sexual enjoyment, namely aphanisis ( Jones 1927) (when first advanced by Jones as an explanation of anxiety, the theory of aphanisis was not related to the anxiety of separation; two years later, however, he sought to adapt it to fit in with Freud's latest ideas); (c) a variant proposed by Spitz ( 1950) and presented within a new theoretical model by Joffe & Sandler ( 1965) is that the traumatic situation to be avoided is one of narcissistic injury. In the history of Freud's thought the signal theory stems from, and is in certain respects the counterpart of, the theory that explains the
281
child's tie to his mother in terms of secondary drive. The variant that regards narcissistic
injury as the trauma threatened also stems from the secondary-drive tradition.
4. Separation anxiety results from a young child's believing when his mother disappears that he has eaten her up or otherwise destroyed her, and that in consequence he has lost her for ever. That belief, it is held, arises from the ambivalent feelings a child has for his mother, an ambivalence made inevitable by the existence within him of a death instinct. Advanced by Melanie Klein ( 1935), the theory can be called, following her terminology,
that of 'depressive anxiety'.
5. As a result of projecting his aggression, a young child perceives his mother as
persecutory, and this leads him to interpret her departure as due to her being angry with him or wishing to punish him. For this reason, whenever his mother leaves him he believes she will either never return or do so only in a hostile mood, and he therefore experiences anxiety. Again following Melanie Klein ( 1934), this can be termed the theory of 'persecutory anxiety'.
6. Initially the anxiety is a primary response not reducible to other terms and due simply to the rupture of a child's attachment to his mother. This can be called the theory of 'frustrated attachment'. It is the counterpart of theories that regard a child's pleasure in his mother's presence as being as primary as his pleasure in food and warmth. A theory of this sort has been
-376-
advanced by James ( 1890), Suttie ( 1935), and Hermann ( 1936), but has never been given much attention in psychoanalytic circles. It is a theory of this type that I advanced in an earlier paper ( Bowlby 1960a) linked to yet another variant of the signal theory. The theory advanced in this work (Chapter 12) is also a combination of the sixth and the third types. It regards separation of a young child from an attachment figure as in itself distressing and also as providing a condition in which intense fear is readily aroused. As a result, when a child senses any further prospect of separation some measure of anxiety is aroused in him.
In Chapter 5 of this volume attention is drawn to the fact that almost all psychoanalytic theorizing about anxiety and fear is conceived in terms of a biological paradigm that antedates modern evolution theory. This accounts, it is believed, for the numerous competing, complex, and contradictory theories to be found in the literature.
Views of main contributors Sigmund Freud
We have seen that it was not until 1926, when Freud was seventy, that in Inhibitions, Symptoms and Anxiety he gave systematic attention to separation anxiety. Prior to this, having paid insufficient attention to the child's attachment to his mother, as he himself affirms ( Freud 1931), he had paid correspondingly little to the anxiety exhibited on separation from her. Nevertheless, he had been far from blind to it. In both the Three Essays ( 1905b) and the Introductory Lectures ( 1917b) he had drawn attention to it and in both had treated it as of much importance. 1
In Three Essays, after a section concerned with early object relations, he gives a paragraph to 'infantile anxiety' ( SE 7:224). In it he advances the view that 'anxiety in children is originally nothing other than an expression of the fact that they are feeling the loss of the person they
282
love'. This view he readily aligns with his hypothesis regarding neurotic anxiety in adults. At that time Freud still held the view that, when a powerful
____________________
1 Separation of child from mother as a central and recurrent theme in Freud's thinking about
anxiety is clearly brought out in Strachey's valuable introduction to the Standard Edition of Inhibitions, Symptoms and Anxiety ( Strachey 1959).
-377-
sexual excitation is insufficiently discharged, libido is transformed directly into anxiety. It is the same in children, he believes. Because 'children . . . behave from an early age as though their dependence on the people looking after them were in the nature of sexual love', and because in a separation situation the child's libido goes unsatisfied, Freud concludes that a child deals with the situation just as an adult would, namely 'by turning his libido into anxiety'. Four years later this is also his explanation of the separation anxiety that was Little Hans's first symptom: 'It was this increased affection for his mother which turned suddenly into anxiety . . . ' ( SE 10: 25).
He follows the same reasoning in the Introductory Lectures ( 1917b). After once again drawing attention to the anxiety exhibited when mother is missing, he concludes that 'infantile anxiety has very little to do with realistic anxiety, but, on the other hand, is closely related to the neurotic anxiety of adults. Like the latter, it is derived from unemployed libido . . . ' ( SE 16: 408). This, it will be observed, is tantamount to identifying neurotic anxiety of adults with separation anxiety of infants, a resemblance on which he had already remarked in 1905. 1
Although in the Introductory Lectures, for reasons which appear inadequate, Freud complicates his theory by postulating that the core of anxiety is a repetition of the affect experienced at birth ( SE 16: 396), it is none the less anxiety arising on separation from mother, as observed empirically, which throughout his writings on infantile anxiety from 1905 onwards holds the centre of the theoretical stage. Anxiety arising at birth, which had first been postulated some years earlier ( 1910, SE II: 173), starts by being only a rather speculative addition to his theory. Although it gradually acquires an equal status, it never usurps the place of anxiety arising on separation from mother. This is important since more than one analyst has tended to give it precedence in his theorizing. 2
____________________
1 '. . . an adult who has become neurotic owing to his libido being unsatisfied behaves in his
anxiety like a child: he begins to be frightened when he is alone . . . and he seeks to
assuage this fear by the most childish measures' ( SE 7: 224).
2 In Introductory Lectures Freud describes a child missing 'the sight of a familiar and
beloved figure -- ultimately of his mother' as the 'situation which is the prototype of the anxiety of children' ( SE 16: 407). However, he thinks that in this situation there may be in addition a reproduction of birth anxiety. In Inhibitions, Symptoms and Anxiety, on the other hand, it
-378-
The next reference to separation anxiety occurs in Beyond the Pleasure Principle ( 1920) where Freud relates the well-known cotton-reel incident which Jones ( 1957: 288) tells us he
283
had witnessed five years previously in Hamburg. His eighteenmonth-old grandson took all sorts of small objects and threw them away into corners and under the bed with an expression which seemed to signify 'gone'. This appeared to be confirmed when later the boy had a cotton reel on the end of a string and played the double game of throwing it away with an expression of 'gone' and pulling it back again with a joyful 'da'. This simple game, coupled with the fact that the boy 'was greatly attached to his mother', led Freud to an
interpretation of the game . . . it was related to the child's great cultural achievement -- the instinctual renunciation (that is, the renunciation of instinctual satisfaction) which he had made in allowing his mother to go away without protesting. He compensated himself for this, as it were, by himself staging the disappearance and return of the objects within his reach ( SE 118: 14-15).
How well established this cultural achievement was we shall never know, but if Freud's grandson followed a common course of development it is unlikely to have been maintained. There are many infants who are able to permit their mother to leave them for an hour or so without crying when they are eighteen months old, but who in the succeeding months find this less tolerable and may make a great fuss. However that may be, the observation of the incident, and no doubt of others like it, seems to have clarified Freud's perception of the child's
____________________
is birth anxiety that is described as the prototype. Nevertheless, in one of the addenda to this work he explains how he could make no headway with Rank's ideas on the primary role of birth trauma and, referring to his own conclusions, remarks that the significance of birth is 'reduced to this prototypic relationship to danger' ( SE 20: 162). This is also the position he takes in the New Introductory Lectures ( 1933)where he repeats his view that 'a particular determinant of anxiety (that is, situation of danger) is allotted to every age of development as being appropriate to it' ( SE 22: 88). The danger situation of birth and the danger of loss of object or of love seem here to be assigned equal status. See also the discussions by Jones ( 1957: 274-6) and by Strachey ( 1959: 83-6). Strachey points out that, in Freud's later work, it is only the form taken by anxiety that is to be understood as stemming from the experience of birth.
-379-
tie to his mother and to have led him to reflect further on the theory of anxiety -- an early example of the value of direct observation.
It was the publication of Rank Trauma of Birth in 1924, Freud relates in an addendum to Inhibitions, Symptoms and Anxiety ( 1926a), that 'obliged me to review the problem of anxiety once more'. In this work Rank had taken up the suggestion which, as we have seen, had first been thrown out by Freud, 'that the affect of anxiety is a consequence of the event of birth and a repetition of the situation then experienced . . . But', Freud continues, 'I could make no headway with his idea that birth is a trauma, states of anxiety a reaction of discharge to it and all subsequent affects of anxiety an attempt to "abreact" it more and more completely' ( SE 20: 161). Instead, what Freud does in his courageous re-examination of theory is to return to the safe ground of empirical observation -- which brings him back once more to separation anxiety.
284
In reading Inhibitions, Symptoms and Anxiety we find that Freud wrestled with the theoretical problems of anxiety through seven chapters, in the course of which he abandons a favourite hypothesis, namely that anxiety represents a direct transformation of libido. His reason for doing so lies in his recognition that, whereas formerly he had supposed anxiety to be the product of repression, an examination of clinical material suggests that, on the contrary, repression is a consequence of anxiety ( SE 20: 109). As a result of this, at the beginning of the eighth chapter he concludes ruefully: 'Up till now we have arrived at nothing but contradictory views about it [anxiety]. . . . I therefore propose to adopt a different procedure. I propose to assemble, quite impartially, all the facts that we do know about anxiety without expecting to arrive at a fresh synthesis' (p. 132 ). After a brief diversion he proceeds:
Only a few of the manifestations of anxiety in children are comprehensible to us, and we must confine our attention to them. They occur, for instance, when a child is alone, or in the dark, or when it finds itself with an unknown person instead of one to whom it is used -- such as its mother. These three instances can be reduced to a single condition, namely that of missing someone who is loved and longed for. But here, I think, we have the key to an understanding of anxiety . . . anxiety appears as a reaction to the felt loss of the object (pp. 136 -7).
-380-
Up to this point Freud is working from empirical data, data moreover which are now amply confirmed. Nevertheless he still remains puzzled, as others have also been, as to how to explain his observations. Why should there be this reaction of anxiety? It 'has all the appearance', he remarks, 'of being an expression of the child's feeling at its wits' end, as though in its still very undeveloped state it did not know how better to cope with its cathexis of longing' (p. 137 ). Today we can draw on a more sophisticated theory of instinctive behaviour to frame a hypothesis which regards the 'cathexis of longing' as the essence of the problem. Fifty years ago, however, such ideas on instinctive behaviour were unknown; instead, Freud was under the impression that the child's attachment could be understood only in terms of secondary drive and that the only primary needs are those of the body.
Freud therefore proceeds:
'The reason why the infant in arms wants to perceive the presence of its mother is only because it already knows by experience that she satisfies all its needs without delay. The situation then, which it regards as a 'danger' and against which it wants to be safeguarded is that of non-satisfaction, of a growing tension due to need, against which it is helpless. '
This, he continues, is 'analogous to the experience of being born . . . What both situations have in common is an economic disturbance caused by an accumulation of amount. 3 of stimulation which require to be disposed of. It is this factor, then, which is the real essence of the "danger" . . . ' and which he terms the 'traumatic situation'. To avoid this, an infant, by a process of learning, displaces 'the danger it fears . . . from the economic situation on to the condition which determined that situation, viz. the loss of object. It is the absence of the mother that is now the danger; and as soon as that danger arises the infant gives the signal of anxiety, before the dreaded economic situation has set in' (pp. 137 -8).
In considering Freud's every approach to the problem of anxiety it is necessary constantly to bear in mind that, from the earliest days of his psychoanalytic theorizing onwards, he adopts as his basic postulate that the nervous system has the function of getting rid of stimuli and that
285
the greatest catastrophe that can befall it is that of being overwhelmed by stimuli (see this volume, Chapter 5). Such theorizing constitutes what Freud describes as the economic viewpoint, and is cast sometimes in
-381-
terms of a psychical energy that builds up and is either discharged in action or else becomes dammed up, and sometimes in terms of excitation or stimulation that similarly varies in quantity. The 'dreaded economic situation' that Freud believes threatens an infant who is separated from his mother is none other than the damming up of psychical energy that cannot be discharged.
As a consequence of his re-examination of the problem, Freud concludes that anxiety has two sources. Anxiety from the first source arises as 'an automatic phenomenon', with physiological features that he believes may well be part of a response appropriate to the situation of birth. Such anxiety occurs whenever a traumatic situation 'is established in the id', that is in 'situation[s] of non-satisfaction in which the amounts of stimulation rise to an unpleasurable height without its being possible for them to be mastered psychically or discharged . . . ' ( SE 20: 137-41). Such traumatic situations are always characterized by helplessness. In Freud's formulation of this source of anxiety we see a direct descendant of his earliest theory, that advanced in his paper on the 'Anxiety Neurosis' ( 1895), in which he postulated that anxiety is developed when the nervous system is incapable of dealing with a mass of excitation.
Anxiety from the second source, Freud suggests, constitutes 'a rescue signal' designed to indicate that danger is impending. Since it requires foresight, such anxiety can 'only be felt by the ego' ( SE 20: 140). It is indeed the task of the ego so to imagine the danger situation in advance that it can restrict 'that distressing experience to a mere indication, a signal' (p. 162 ). Freud proceeds to list a number of danger situations, each corresponding to a particular developmental phase, which, if allowed to develop, would result in a traumatic situation: among these are birth, loss of object (namely mother), fear of father, and fear of superego (pp. 146 -7).
In his account of this second source of anxiety Freud lays much emphasis on the elements of foresight and expectation: 'The individual will have made an important advance in his capacity for self-preservation if he can foresee and expect a traumatic situation . . . which entails helplessness, instead of simply waiting for it to happen' (p. 166 ). Before anxiety can arise from this source, therefore, a fair degree of cognitive development is necessary.
Although, as has been indicated, Freud conceives of separation anxiety itself as no more than a signal and as being
-382-
developed through a process of learning, as indeed is necessary if it is based on foresight, it is evident that he is not entirely satisfied with that conclusion. At the end of the book ( SE 20: 168) he returns yet again to 'the puzzling phobias of early childhood' and hazards that perhaps, as in other species, the fear of loss of object may be a built-in response: thus he refers to an 'archaic heritage' and 'vestigial traces of the congenital preparedness to meet real dangers'. These reflections, and also the similar ones in regard to the child's tie that are found in the Outline ( 1940), which were noted in the appendix to the first volume of the present work,
286
suggest that towards the end of his life Freud was moving towards a formulation not very different from that advanced here.
It is far more important, however, that in this late work Freud finally clarifies what is held here to be the true relatedness of separation anxiety to mourning and defence. Previously, as he candidly admits, he had been confused. Not only had he supposed that repression is antecedent to anxiety, but he had also found it difficult to believe that anxiety as well as grief can be a response to loss of object. Now he sees that sequence clearly: anxiety is the reaction to the danger of loss of object, the pain of mourning is the reaction to the actual loss of object, and defences protect the ego against instinctual demands which threaten to overwhelm it and which can occur all too readily in the absence of the object ( SE 20: 164-72). This formula has not commonly been adopted by later theorists.
Ernest Jones
When Jones ( 1927) first advanced his theory of aphanisis it is evident, from the absence of references to Inhibitions, Symptoms and Anxiety, that he was unaware of Freud's latest train of thought. Furthermore, there is evidence also that he was still unaware of the importance of the attachment to mother (irrespective of the child's sex). It must therefore be noted that Jones's theory of aphanisis -- that 'the fundamental fear is [of] the total and of course permanent extinction of the capacity (including opportunity) for sexual enjoyment' -- was advanced without any reference to the present topic. The only mention of separation anxiety is in reference to weaning as a pregenital precursor of castration and to a girl's fear of separation from her father.
Two years later, however, Jones ( 1929) strives to integrate his own theory of aphanisis with Freud's theory of signal
-383-
anxiety. The union is uneasy and the resulting theory appreciably more complex than either taken singly. One of several difficulties is that Jones is still not aware of the tie to the mother irrespective of the child's sex. 1 Since the combined theory has been little called upon, to delineate it is unnecessary. As a broad generalization, it may be said that Jones accepts Freud's view regarding signal anxiety, believing it to be 'purposely provoked by the ego so as to warn the personality' of the possible approach of serious dangers, and then, in describing these serious dangers, adds to Freud's conception of what constitutes the 'traumatic situation' his own notion of aphanisis.
Melanie Klein
Whereas Jones developed his theory of anxiety independently of Freud's and later attempted to marry the two, Melanie Klein not only developed hers independently of Freud's but has frequently underlined the differences between them. Anxiety, in her judgement, is to be understood in terms of the death instinct, to which Freud never referred in this connection, and therefore in terms of aggression. Her views in regard to anxiety in general, which were taking shape between 1924 and 1934, and to separation anxiety in particular, are set out fully in her paper 'On the Theory of Anxiety and Guilt' ( 1948b, in Klein et al. 1952). They represent the only formulation made by a psychoanalyst which is both substantially different
287
from the formulations discussed by Freud and has had significant influence on theory and practice.
In Inhibitions, Symptoms and Anxiety, following a line of argument already advanced in Introductory Lectures ( SE 16: 407-8), Freud explicitly rejected the notion that fear of death is a primary anxiety and concluded instead that it is a later and learnt fear. 2 Melanie Klein differs: 'I do not share this view because my analytic observations show that there is in the unconscious a fear of annihilation of life. ' This she assumes must
____________________
1 See, for instance, his reference to the external danger arising from withdrawal of object,
'e. g. the mother in the boy's case' (p. 311, my italics).
2 It is interesting to note that Sylvia Anthony ( 1940) in her study of the genesis of children's
ideas of death reached a similar conclusion. Furthermore, she believes that it is through its equation with separation that death acquires its emotional significance: 'Death is equated with departure. . . . To the young child death means, in the departure context, its mother's death -- not its own. '
-384-
be the response to the death instinct: 'Thus in my view the danger arising from the inner working of the death instinct is the first cause of anxiety' ( Klein et al. 1952: 276). This, she suggests, is felt by an infant 'as an overwhelming attack, as persecution', and a persecution, moreover, which is first experienced at birth: 'We may assume that the struggle between life and death instincts already operates during birth and accentuates the persecutory anxiety aroused by this painful experience. ' From this argument she draws an important conclusion regarding the infant's first object relations: 'It would seem', she says, 'that this experience [i. e. birth] has the effect of making the external world, including the first external object, the mother's breast, appear hostile' ( 1952: 278). In another paper ( 1946) she summarizes her view in a sentence: 'I hold that anxiety arises from the operation of the death instinct within the organism, is felt as fear of annihilation (death) and takes the form of fear of persecution' ( Klein et al. 1952: 296). It is against this backcloth -- that anxiety is the result of the perpetual activity of the death instinct and that the newborn infant is already burdened with persecutory anxiety -- that Klein presents her views on separation anxiety.
Starting from Freud's distinction between objective anxiety (arising in connection with a known external danger) and neurotic anxiety (arising in connection with an unknown and internal one) ( Freud 1926a, SE 20: 165 and 167), Klein (1948b) sees both as contributing to the infant's fear of loss. She describes their nature as follows: objective anxiety arises from 'the child's complete dependence on the mother for the satisfaction of his needs and the relief of tension'; neurotic anxiety 'derives from the infant's apprehension that the loved mother has been destroyed by his sadistic impulses or is in danger of being destroyed, and this fear . . .
contributes to the infant's feeling that she will never return'. Were Klein to postulate that this depressive anxiety only developed in later infancy, she would not be diverging materially from Freud's view but only expanding it at an important point. This, however, is not her position. She emphasizes that in her view both sources of anxiety are present from the beginning and are constantly interacting. Because of this, 'no danger-situation arising from external sources could ever be experienced by the young child as a purely external and known danger' ( Klein et al. 1952: 288). On this her own statements and those of her colleagues are consistent. In discussing the cotton-reel incident Klein dissociates
288
-385-
herself explicitly from Freud's view and concludes, 'when [an infant] misses [his mother], and his needs are not satisfied her absence is felt to be the result of his destructive impulses' (pp. 269 -70). In the same volume it is claimed by Susan Isaacs that always 'mental pain has a content, a meaning, and implies phantasy. On the view presented here, "he behaves as if he were never going to see her again" 1 means his phantasy is that his mother has been destroyed by his own hate or greed and altogether lost' ( Klein et al. 1952: 87).
These passages seem to make it clear that in their explanations of separation anxiety Melanie Klein and her colleagues see depressive anxiety as virtually its sole component. This, however, is not so since elsewhere they emphasize that the relationship to the mother is itself 'a first measure of defence . . . The dependence on the mother and fear of loss of her, which Freud regards as the deepest source of anxiety, is from our point of view (the self- preservative) already a defence against a greater danger (that of helplessness against destruction within)' ( Joan Riviere in Klein et al. 1952: 46-7). 'From the very beginning,' she writes, 'the internal forces of the death instinct and of aggression are felt to be the cardinal danger threatening the organism' (p. 44 ). Since these forces are let loose during a separation experience, in the final analysis separation anxiety is seen as a response to the threat of destruction within. Clearly, this theory is very different from that of Freud and also from that advanced here. Whereas Freud gives primacy to anxiety that arises from 'an accumulation of amounts of stimulation' that he conceives as resulting from separation, Melanie Klein and her colleagues give primacy to persecutory anxiety.
It should, however, be added that in various passages Klein refers also to birth as constituting an anxiety-provoking trauma, and seems at times to subscribe to the birth-trauma theory of separation anxiety. Thus, following a passage already quoted above ( 1952: 296), she writes: 'Other important sources of primary anxiety are the trauma of birth (separation anxiety) and frustration of bodily needs. ' Nevertheless, although postulating these additional sources of anxiety, she quickly brings them within the ambit of persecutory anxiety by attributing to an infant a tendency always to suppose fear to be aroused by an object. After having earlier expressed the opinion that 'the fear of the destructive impulse seems to attach itself at once to
____________________
1 Isaacs quotation, from Inhibitions, Symptoms and Anxiety, is from the 1936 English
translation ( London: Hogarth, p. 167). -386-
an object', she completes her statement regarding the trauma of birth and the frustration of bodily needs thus: 'and these experiences too are from the beginning felt as being caused by objects. 1 Even if these objects are felt to be external, they become through introjection internal persecutors and thus reinforce the fear of the destructive impulse within' (p. 296 ).
In evaluating Melanie Klein's views it is essential to realize that her main theoretical outlook was formed in the years preceding the publication of Freud Inhibitions, Symptoms and Anxiety, and that, unlike Freud, who in the final formulation of his theory took anxiety arising from separation experiences as his point of departure, Klein had already developed her theory of anxiety before she gave any attention to separation from mother as a situation that provokes
289
anxiety. The first occasion she discusses it is in 1935 in her paper on 'The Psychogenesis of Manic-depressive States'.
When we look back on the early papers of Melanie Klein we remain impressed by her observation that anxiety and unconscious aggression often coexist, particularly when there is an unusually anxious and intense attachment of one person to another. In my judgement, however, she assumed too readily that aggression both precedes and causes anxiety so that, instead of recognizing conscious and unconscious aggression as a common response to separation and as constituting an important and frequent condition for the exacerbation of separation anxiety, she came to see aggression as the single source of anxiety; and furthermore, by identifying the child's tie to his mother with orality, was led into making implausible assumptions about the mental life of infants during their early months and thence into creating a theoretical superstructure that is far from convincing. This has had two unfortunate results. On the one hand, some of her critics have failed to appreciate the value of certain parts of her contribution; on the other, her followers have been slow to recognize that, significant though depressive and persecutory anxieties may sometimes be, the origin of separation anxiety cannot be understood in such terms, and, more important, that disturbances of the mother--child relationship that arise during the second and many subsequent years can have a far-reaching potential for pathological development.
____________________
1 Freud is not favourable to this type of theory. He writes: 'A child who is mistrustful in this
way and terrified of the aggressive instinct which dominates the world is a theoretical construction that has quite miscarried' ( SE 16: 407).
-387-
Anna Freud
Whereas Melanie Klein has written much about separation anxiety but has recounted few observations of how infants and young children actually behave in situations of separation, Anna Freud was one of the first to record such observations but until recent years has discussed their theoretical implications singularly little. As in the case of Klein, it looks as though a main reason was that her theoretical orientation was already set before Freud's fresh appraisal of the nature and genesis of anxiety appeared. Inhibitions, Symptoms and Anxiety is not referred to in her book The Psycho-analytical Treatment of Children ( 1946), which dates from 1926, 1927, and 1945; and, though a chapter is given to processes of defence in relation to the source of anxiety or to danger, there is no reference in The Ego and Mechanisms of Defence ( 1936) either to separation anxiety or to loss of object. Until her experiences with babies and young children in the Hampstead Nurseries during the war, Anna Freud seems to have given little attention to these problems.
In the two modest volumes published with Dorothy Burlingham ( Burlingham & Freud 1942; 1944), observation is sharp and description telling. Of children aged between one and three years they write: 'Reactions to parting at this time of life are particularly violent . . . This new ability to love finds itself deprived of the accustomed objects and his greed for affection remains unsatisfied. His longing for his mother becomes intolerable and throws him into states of despair' ( 1942:51). Yet, despite this clear understanding of the distress that is implicit in these responses, neither in these two volumes nor in the papers published by Anna
290
Freud during the subsequent decade are such manifestations related in any systematic way to anxiety in general or to separation anxiety in particular.
Instead, one has the impression that Burlingham & Freud were unprepared for the intensity of the responses they saw in the Nurseries and puzzled how to explain them. For instance, there is a passage ( 1942: 75-7) where they express the belief that perhaps if separations could be arranged more gradually all would be well: 'It is not so much the fact of separation to which the child reacts as the form in which the separation has taken place. ' In another passage (p. 57 ) the distress of a child between three years and five years seems to be attributed, entirely, to his belief that separation is a punishment--'To overcome this guilt he overstresses all the love which he has ever
-388-
felt for his parents'--a comment which suggests that in their view there would be no distress at this age were there not guilt and persecutory anxiety. Perhaps they get nearer the truth when in these same passages they refer to 'the natural pain of separation' and to the fact that 'unsatisfied longing produces in him a state of tension which is felt as shock'.
Whenever during that period Anna Freud broaches a theoretical interpretation of these responses or of the long-term results of separation (e. g. 1952; 1953), she takes for granted that the child's tie to his mother is to be accounted for by the theory of secondary drive. Since the infant has no needs but those of his body, his interest is at first confined to anyone who meets those needs; in so far as there is anxiety at separation from mother, it is a result of the fear that bodily needs will go unmet. Her views are perhaps most clearly expressed in an address to medical students ( 1953). After describing her conception of how attachment grows in the well-cared-for child she proceeds:
On the other hand, in cases where the mother has carried out her job as provider indifferently, or has allowed too many other people to substitute for herself, the transformation from greedy stomach-love to a truly constant love attachment is slow to come. The infant may remain too insecure and too worried about the fulfilment of his needs to have sufficient feeling to spare for the person or persons who provided for them [my italics].
This conclusion is a logical outcome of the secondary-drive theory of the child's tie and of Freud's version of the signalanxiety theory of separation anxiety.
More recently, in a book published in 1965, Anna Freud describes several 'forms' taken by anxiety during the early years, each of which she believes to be characteristic of a particular phase in the development of object relations. The sequence of forms runs as follows: 'archaic fears of annihilation, . . . separation anxiety, castration anxiety, fear of loss of love, guilt . . . ' Separation anxiety (and also fear of annihilation) is held to be characteristic of, and confined to, the first phase in the development of object relations; that is described as a phase of 'biological unity between the mother--infant couple, with the mother's narcissism extending to the child, and the child including the mother in his internal "narcissistic milieu" . . . '. During subsequent phases, forms of anxiety other than separation anxiety are thought to occur. For example, the third
-389-
291
phase, described as that of object constancy, is believed to be characterized by fear of the loss of the object's love. ?
Contributions by Other Exponents of Ego Psychology
The theories advanced by Anna Freud during earlier years are subscribed to also by Nunberg ( 1932), Fenichel ( 1945), and Schur ( 1953; 1958). In his two carefully reasoned studies of anxiety, Schur makes the commonly held assumption that in man the biologically given components of behaviour are strictly limited. In the later paper, in which he draws extensively on ethological data and concepts, he details what he believes they comprise. On the one hand, he postulates the presence of fight and flight reactions characteristic of the phase of development which begins with the ability to perceive external objects. On the other, he postulates an earlier phase ('the undifferentiated phase') during which 'all danger is, due to the infantile development specific for man, "economical", inner danger', namely danger arising from an accumulation of excitation that springs from unmet bodily needs. It is from this, specifically human, source of anxiety that he regards separation anxiety as developing as a learnt derivative: 'The realization that an external object can initiate or end a traumatic situation displaces the danger from the economic situation to the condition which determines that situation. Then it is no longer hunger that constitutes danger for the child but it is the absence of the mother. ' Although he discusses various dangers which he thinks 'may be based on innate givens', nowhere does he consider the possibility that loss of mother may be one of them.
After he has come to recognize the importance of separation anxiety Kris ( 1950) makes a serious effort to incorporate it in his theorizing. But his views are based more on inference from previous theory than on a reassessment of the data; in particular he is concerned, like Schur, to cast them in a form compatible with Hartmann's ego psychology. This leads him to place great emphasis on a distinction between the danger of losing the love object and the danger of losing the object's love. Although this distinction was referred to briefly by Freud ( 1926a), the way that Kris elaborates it is his own. On theoretical grounds he postulates that the danger of losing the love object is concerned solely with anaclitic (namely bodily) needs and is not concerned with a particular love object. Conversely, the development of a 'relationship to a permanent personalized love object that can no longer easily be replaced' he postulates
-390-
to occur synchronously with the development of responsiveness to the danger of losing the object's love; it represents, in his opinion, 'a decisive step in ego development'.
This hypothetical association is not, however, borne out by observation. Anxiety reactions to the loss of a particular love object are to be seen some months before it is reasonable to credit a human infant with awareness of the danger of losing the object's love and before the twelve- month age-limit suggested by Kris ( 1950). As is emphasized in the previous volume (Chapter 15), the responses mediating attachment behaviour both in man and in lower species tend quickly to focus on a particular figure; and there can be no reason to suppose that their doing so represents an important step in ego development. In the event, therefore, the theoretical distinction advanced by Kris must be regarded as mistaken.
292
The crucial connection between anxiety as the reaction to the danger of losing the object and the pain of mourning as the reaction to its actual loss, which Freud arrives at in the final pages of Inhibitions, Symptoms and Anxiety, has been little recognized. Only in the work of Melanie Klein and Therese Benedek is it given much place. Helene Deutsch ( 1937) explicitly divorces the two: anxiety is an infantile response, she holds, grief and mourning more mature ones. 'The early infantile anxiety', she writes, 'we know as the small child's reaction to separation from the protecting and loving person. ' When the child is older, on the other hand, 'suffering and grief [are] to be expected in place of anxiety' (p. 14, my italics). Moreover, separation anxiety in the older individual is to be understood always as a regression to infancy, and occurs in situations where 'grief . . . threaten[s] the integrity of the ego, or, in other words, if the ego [is] too weak to undertake . . . mourning' (p. 14 ). This differentiation by maturity does not stand examination, however. In the responses of infants and young children to loss of mother, elements of grief are undoubtedly present. Conversely, as Therese Benedek among others has recorded, anxiety is the rule even in adults when they are separated for any length of time from someone they love.
For many years Therese Benedek has been concerned with problems of separation from, and reunion with, loved persons, and with responses to loss and bereavement; and as a result of her clinical work she has had a lively awareness of the farreaching significance of separation anxiety and of its close relatedness to anxiety and mourning. In describing responses to
-391-
separations, reunions, and bereavements occurring during wartime, she frequently speaks of separation as a trauma in itself, and she generalizes boldly: 'The universal response to separation is anxiety' ( Benedek 1946: 146). She also recognizes that the experience of being separated, or the expectation of being separated, from a loved person leads to a sharp increase in longing for his company. In a later paper ( 1956) she notes that a crying fit in an infant is by no means always caused 'by a commanding physiologic need such as hunger and pain, but by the thwarting of an attempt at emotional (psychologic) communication and satisfaction'.
All these observations can be parsimoniously explained in terms of the theories regarding attachment, separation anxiety, grief and mourning that are advanced in the present work. Nevertheless, although her original training took place in Budapest (see Appendix to Volume I), Benedek does not accept these simpler hypotheses. Instead, in all her theorizing she is committed to a secondary-drive theory of the child's tie to his mother with all its complications and disadvantages. Thus the increase in longing evident in adults at separation, which can hardly be considered other than a natural and normal response, is explained as due to a regression to oral dependency. Indeed, as in so much theorizing deriving from the concept of dependence, Therese Benedek tends at times to theorize as though all attachments to loved persons were undesirable regressions to an infantile state.
Nowhere in Benedek's writings is there any systematic discussion of separation anxiety; but, in the later paper ( 1956) referred to above, two separate theories appear to be adumbrated. The first is similar to Freud's signal-anxiety theory; the second is concerned with the danger of ego disorganization.
Still struggling with the same problem that Freud was wrestling with thirty years earlier, she asks why an infant should respond to 'the frustration of a "dependent" wish' by crying. Reverting to the belief that crying is related intrinsically only to the experiences of hunger and
293
pain, she concludes that 'he responds to the lack of participation from the adult as to a complete interruption of the symbiosis, as if he were abandoned and hungry' (p. 402, my italics).
Since, however, she is not altogether confident that crying is to be understood as anxiety, and believes that anxiety proper is a response to the danger of ego distintegration, Benedek advances another view. This is that the young child has to turn
-392-
to his mother to preserve his ego integration when faced with the 'anxiety, humiliation and shame of failure'. In the case of the older child, 'his ego can maintain itself by its own resources' (pp. 408 -9). Thus, although the clinical data she presents are consistent with the theory advanced here, Benedek's interpretations remain firmly embedded within the traditional paradigm.
In much of her theorizing, especially in her use of the concept of symbiosis, Margaret Mahler ( 1968) follows Therese Benedek; and in attributing a distinct form of anxiety to each phase of the development of object relations she follows Anna Freud. Nevertheless, despite the similarity of their postulated phases, the phase of development to which Mahler attributes separation anxiety is not the same as the one to which Anna Freud attributes it. Whereas Anna Freud regards separation anxiety as a response specific to 'infringements of the biological mother-infant tie' during the first phase of development, Mahler holds that separation anxiety is attributable properly only to a later phase, namely the phase 'after the beginning of object constancy has been achieved'. This she puts in the third and fourth years. The form of anxiety that Mahler attributes to the first phase of development, the symbiotic phase, is a fear of self- annihilation, the reasoning being that at that phase 'loss of the symbiotic object' is thought to amount 'to loss of an integral part of the ego itself'. This mode of theorizing is close to that of Spitz.
Although, like most other analysts, Spitz is an adherent of the secondary-drive theory to account for the child's tie to his mother and endorses Freud's version of the signal-anxiety theory of separation anxiety ( 1950), he advances, in addition, a variant of that theory. This is a theory of 'narcissistic trauma'. After outlining his views on the development of object relations from a phase of narcissism (first three months) through a phase of pre-objectal relations (second three months) to a phase of true object relations (third three months), he proceeds:
It is in the third quarter that true objects appear for the first time. They now have a face, but they still retain their function of a constituent part of the child's recently established Ego. The loss of the object is therefore a diminution of the Ego at this age and is as severe a narcissistic trauma as a loss of a large part of the body. The reaction to it is just as severe.
From other passages, in which he insists on the warning function
-393-
of anxiety and its dependence on learning and foresight, it is clear that in Spitz's view anxiety is a signal to warn against the danger of a narcissistic trauma. This is a fresh variant of the
294
signal-anxiety theory: the traumatic situation to be avoided is, this time, one in which narcissism is threatened.
It should be noted that much of Spitz's theorizing about anxiety turns around his concern to explain the anxiety exhibited by an infant of seven or eight months when confronted by a stranger, which he was the first to term eight-months' anxiety; the anxiety exhibited at separation from a loved object is less in his mind. In view of his empirical work this may seem surprising, until we realize that his observations of deprived infants were not concerned with the immediate responses to separation, namely protest, distress, and anxiety, or with responses after reunion, but were largely concentrated on responses seen during the later phases of separation, namely grief and depression. As a result he had no opportunity to observe the continuum of response from separation anxiety to grief and mourning.
The approach of Sandler & Joffe to these problems follows fairly closely the approaches of Kris and Spitz. The traditional theory of secondary drive is adopted to account for the child's tie to his mother, together with the concept of dependency. In keeping with their basic model, moreover, they place almost exclusive emphasis on the feeling states produced in a child by the presence or absence of his mother and make little attempt to relate these feeling states either to instinctive behaviour or to the survival value of mother's presence and the increased risk attendant on her absence. Thus they describe 'the role of the object' in a child's life as 'that of a vehicle for the attainment of the ideal state of well-being'. Conversely, 'loss of the object signifies the loss of an aspect of the self', i. e. that 'part of the self-presentation which . . . reflects the relation to the object' ( Joffe & Sandler 1965, their italics).
For Sandler & Joffe, therefore, as for Freud, the situation at all costs to be avoided is not so much the actual loss as the traumatic overwhelming of the ego to which loss leads. In terms of the Sandler & Joffe model, the traumatic situation to be avoided is described as a 'disruption of the individual's feeling state' ( Sandler & Joffe 1969).
Summing up the differences they see between their theoretical interpretation and my own, Joffe & Sandler ( 1965) conclude:
-394-
Object loss may bring about acute mental pain through creating a wound in the self. This view coincides with what Abraham and others have described as the 'severe injury to infantile narcissism' which object loss entails. And although Bowlby has maintained ( 1960b) that such a statement misses the true significance of object loss, we take the view that it contains its essence.
Such contrasting positions are, of course, a simple consequence of our having adopted different paradigms.
Other Contributors
In view of Sullivan's insistence that psychiatry is the study of interpersonal relationships, it is not unexpected that he sees all anxiety as a function of the child's relationship to his mother and other significant people. Nevertheless his position is different from that advanced here, especially in the primacy that he gives to the role of learning; for he regards anxiety as being exclusively a product of the mother's attitude. When mother is approving, her child is content;
295
when she is disapproving, her child is anxious. Despite his great emphasis on 'need for contact' and 'need for tenderness' and the strong terms in which he refers to the experience of loneliness -- 'really intimidating' and 'terrible' ( 1953: 261) -- that separation from a loved object can of itself induce anxiety appears to be explicitly ruled out. Thus, in a final chapter, he indicates features that man has in common with other species, namely bodily needs and 'even our recurrent need for contact with others'. These he contrasts with features 'restricted to man and some of the creatures he has domesticated', which include 'the experience of anxiety' (p. 370 ). His assumption that anxiety is confined to domesticated species follows from his assumption that it results from processes of training and learning: 'there is nothing I can conceive in the way of interpersonal action about which one could not be trained to be anxious' (pp. 370 -1). Even 'the experience of intense anxiety' which gives rise to repression is conceived as resulting from ill-conceived educational methods (p. 163 ). 1
Although in Sullivan's view the induction of anxiety remains
____________________
1 Dr Mabel Blake Cohen has emphasized that Sullivan did not regard such 'training' as a
product only of conscious parental attitudes: ' Sullivan recognized that unconscious attitudes or tensions in the parents' interactions with the child were of considerably more importance than conscious planned behaviour' (personal communication).
-395-
something of a mystery -- 'the character of situations which provoke anxiety is never completely to be grasped' (p. 190 ) -- it is nevertheless evident that in effect he sees it as always connected with processes of child-training. Since he believes that a main anxiety- inducing sanction used by a mother is restriction or denial of tenderness (p. 162 ), he comes at times near to my own concept of separation anxiety and its exacerbation by threats of abandonment. What seems to escape Sullivan, however, is that distress and anxiety can be and often are direct consequences of lack of tenderness and of separation per se; and that threats to restrict tenderness would be ineffective were that not so. While aware that loneliness can be a devastating experience for adolescents and adults. Sullivan seems unaware that it is even more distressing for infants and young children; indeed, there are passages in which he seems specifically to exclude that that is so: 'Loneliness, as an experience which has been so terrible that it practically baffles clear recall, is a phenomenon ordinarily encountered only in preadolescence and afterwards' (p. 261, my italics).
Reading Sullivan's work one gets the impression that he had never observed young children and that he was only partially aware of the close attachment they form to particular people and of the sense of security that mere proximity to a loved figure brings. The 'need for contact with others, often felt as loneliness', is identified, not with need for a genital or a parent-child relationship, but with gregariousness in animals (p. 370 ); his conviction that 'no action of the infant is consistently and frequently associated with the relief of anxiety' (p. 42 ), which overlooks the relief an infant commonly exhibits when clutching his mother, is a main plank in his theorizing. Because of this, he seems never to have grasped the reality of separation anxiety and, therefore, despite his close attention to the problems to which it gives rise, it remains almost impossible to attribute to him any particular theory of its nature and origin. It is probably for the same reasons that neither grief nor mourning plays any significant part in his system of psychopathology.
296
In the theorizing of Phyllis Greenacre ( 1952) separation anxiety, and grief and mourning seem also to be omitted. Instead, experiences during the birth process and the first weeks of postnatal life are advanced as major variables to account for a later differential liability to neurosis (see Chapter 16 of this volume).
Rank's views regarding birth trauma have already been
-396-
referred to. In his early papers Fairbairn, who sees separation anxiety as the mainspring of all psychopathology, follows Rank closely in regard to its origins: Fairbairn's postulate ( 1943) that birth anxiety is 'the prototype of all the separation anxiety which is subsequently experienced' is the counterpart of his postulate that a return-to-womb craving accounts for the child's tie. It should be added, however, that these views are peripheral to Fairbairn's main theoretical position ( Fairbairn 1952), which is in all other respects consistent with the theory of frustrated attachment advanced here. In a late paper ( 1963) in which he gives a synopsis of his views he writes: 'The earliest and original form of anxiety, as experienced by the child, is separation anxiety. '
Others have also founded their psychopathology on the central role of separation anxiety and some have adopted a frustrated attachment theory to account for it. For instance, as long ago as 1935, Suttie, holding the view that the child's attachment to his mother is the result of a primary 'need for company', saw anxiety as 'an expression of apprehension of discomfort at the frustration, or threatened frustration, of this all-important motive'. A year later Hermann ( 1936) expressed an almost identical view. He relates anxiety to the urge to seek and cling to mother: 'Anxiety is basically the feeling of being left on one's own in the face of danger. Its expression is a seeking for help and at the same time a seeking for mother.
____________________
1 A version of this review was published in the Journal of Child Psychology and Psychiatry,
Vol. I, 1961. Only few changes have been made; papers published since 1960 are not as systematically considered as those published earlier.
-375-
in many variants according to how the traumatic situation to. be avoided is conceived.
Principal variants are: (a) the traumatic situation is an economic disturbance that is caused when there develops an excessive accumulation of stimulation arising from unsatisfied bodily needs ( Freud 1926a); (b) it is the imminence of a total and permanent extinction of the capacity for sexual enjoyment, namely aphanisis ( Jones 1927) (when first advanced by Jones as an explanation of anxiety, the theory of aphanisis was not related to the anxiety of separation; two years later, however, he sought to adapt it to fit in with Freud's latest ideas); (c) a variant proposed by Spitz ( 1950) and presented within a new theoretical model by Joffe & Sandler ( 1965) is that the traumatic situation to be avoided is one of narcissistic injury. In the history of Freud's thought the signal theory stems from, and is in certain respects the counterpart of, the theory that explains the
281
child's tie to his mother in terms of secondary drive. The variant that regards narcissistic
injury as the trauma threatened also stems from the secondary-drive tradition.
4. Separation anxiety results from a young child's believing when his mother disappears that he has eaten her up or otherwise destroyed her, and that in consequence he has lost her for ever. That belief, it is held, arises from the ambivalent feelings a child has for his mother, an ambivalence made inevitable by the existence within him of a death instinct. Advanced by Melanie Klein ( 1935), the theory can be called, following her terminology,
that of 'depressive anxiety'.
5. As a result of projecting his aggression, a young child perceives his mother as
persecutory, and this leads him to interpret her departure as due to her being angry with him or wishing to punish him. For this reason, whenever his mother leaves him he believes she will either never return or do so only in a hostile mood, and he therefore experiences anxiety. Again following Melanie Klein ( 1934), this can be termed the theory of 'persecutory anxiety'.
6. Initially the anxiety is a primary response not reducible to other terms and due simply to the rupture of a child's attachment to his mother. This can be called the theory of 'frustrated attachment'. It is the counterpart of theories that regard a child's pleasure in his mother's presence as being as primary as his pleasure in food and warmth. A theory of this sort has been
-376-
advanced by James ( 1890), Suttie ( 1935), and Hermann ( 1936), but has never been given much attention in psychoanalytic circles. It is a theory of this type that I advanced in an earlier paper ( Bowlby 1960a) linked to yet another variant of the signal theory. The theory advanced in this work (Chapter 12) is also a combination of the sixth and the third types. It regards separation of a young child from an attachment figure as in itself distressing and also as providing a condition in which intense fear is readily aroused. As a result, when a child senses any further prospect of separation some measure of anxiety is aroused in him.
In Chapter 5 of this volume attention is drawn to the fact that almost all psychoanalytic theorizing about anxiety and fear is conceived in terms of a biological paradigm that antedates modern evolution theory. This accounts, it is believed, for the numerous competing, complex, and contradictory theories to be found in the literature.
Views of main contributors Sigmund Freud
We have seen that it was not until 1926, when Freud was seventy, that in Inhibitions, Symptoms and Anxiety he gave systematic attention to separation anxiety. Prior to this, having paid insufficient attention to the child's attachment to his mother, as he himself affirms ( Freud 1931), he had paid correspondingly little to the anxiety exhibited on separation from her. Nevertheless, he had been far from blind to it. In both the Three Essays ( 1905b) and the Introductory Lectures ( 1917b) he had drawn attention to it and in both had treated it as of much importance. 1
In Three Essays, after a section concerned with early object relations, he gives a paragraph to 'infantile anxiety' ( SE 7:224). In it he advances the view that 'anxiety in children is originally nothing other than an expression of the fact that they are feeling the loss of the person they
282
love'. This view he readily aligns with his hypothesis regarding neurotic anxiety in adults. At that time Freud still held the view that, when a powerful
____________________
1 Separation of child from mother as a central and recurrent theme in Freud's thinking about
anxiety is clearly brought out in Strachey's valuable introduction to the Standard Edition of Inhibitions, Symptoms and Anxiety ( Strachey 1959).
-377-
sexual excitation is insufficiently discharged, libido is transformed directly into anxiety. It is the same in children, he believes. Because 'children . . . behave from an early age as though their dependence on the people looking after them were in the nature of sexual love', and because in a separation situation the child's libido goes unsatisfied, Freud concludes that a child deals with the situation just as an adult would, namely 'by turning his libido into anxiety'. Four years later this is also his explanation of the separation anxiety that was Little Hans's first symptom: 'It was this increased affection for his mother which turned suddenly into anxiety . . . ' ( SE 10: 25).
He follows the same reasoning in the Introductory Lectures ( 1917b). After once again drawing attention to the anxiety exhibited when mother is missing, he concludes that 'infantile anxiety has very little to do with realistic anxiety, but, on the other hand, is closely related to the neurotic anxiety of adults. Like the latter, it is derived from unemployed libido . . . ' ( SE 16: 408). This, it will be observed, is tantamount to identifying neurotic anxiety of adults with separation anxiety of infants, a resemblance on which he had already remarked in 1905. 1
Although in the Introductory Lectures, for reasons which appear inadequate, Freud complicates his theory by postulating that the core of anxiety is a repetition of the affect experienced at birth ( SE 16: 396), it is none the less anxiety arising on separation from mother, as observed empirically, which throughout his writings on infantile anxiety from 1905 onwards holds the centre of the theoretical stage. Anxiety arising at birth, which had first been postulated some years earlier ( 1910, SE II: 173), starts by being only a rather speculative addition to his theory. Although it gradually acquires an equal status, it never usurps the place of anxiety arising on separation from mother. This is important since more than one analyst has tended to give it precedence in his theorizing. 2
____________________
1 '. . . an adult who has become neurotic owing to his libido being unsatisfied behaves in his
anxiety like a child: he begins to be frightened when he is alone . . . and he seeks to
assuage this fear by the most childish measures' ( SE 7: 224).
2 In Introductory Lectures Freud describes a child missing 'the sight of a familiar and
beloved figure -- ultimately of his mother' as the 'situation which is the prototype of the anxiety of children' ( SE 16: 407). However, he thinks that in this situation there may be in addition a reproduction of birth anxiety. In Inhibitions, Symptoms and Anxiety, on the other hand, it
-378-
The next reference to separation anxiety occurs in Beyond the Pleasure Principle ( 1920) where Freud relates the well-known cotton-reel incident which Jones ( 1957: 288) tells us he
283
had witnessed five years previously in Hamburg. His eighteenmonth-old grandson took all sorts of small objects and threw them away into corners and under the bed with an expression which seemed to signify 'gone'. This appeared to be confirmed when later the boy had a cotton reel on the end of a string and played the double game of throwing it away with an expression of 'gone' and pulling it back again with a joyful 'da'. This simple game, coupled with the fact that the boy 'was greatly attached to his mother', led Freud to an
interpretation of the game . . . it was related to the child's great cultural achievement -- the instinctual renunciation (that is, the renunciation of instinctual satisfaction) which he had made in allowing his mother to go away without protesting. He compensated himself for this, as it were, by himself staging the disappearance and return of the objects within his reach ( SE 118: 14-15).
How well established this cultural achievement was we shall never know, but if Freud's grandson followed a common course of development it is unlikely to have been maintained. There are many infants who are able to permit their mother to leave them for an hour or so without crying when they are eighteen months old, but who in the succeeding months find this less tolerable and may make a great fuss. However that may be, the observation of the incident, and no doubt of others like it, seems to have clarified Freud's perception of the child's
____________________
is birth anxiety that is described as the prototype. Nevertheless, in one of the addenda to this work he explains how he could make no headway with Rank's ideas on the primary role of birth trauma and, referring to his own conclusions, remarks that the significance of birth is 'reduced to this prototypic relationship to danger' ( SE 20: 162). This is also the position he takes in the New Introductory Lectures ( 1933)where he repeats his view that 'a particular determinant of anxiety (that is, situation of danger) is allotted to every age of development as being appropriate to it' ( SE 22: 88). The danger situation of birth and the danger of loss of object or of love seem here to be assigned equal status. See also the discussions by Jones ( 1957: 274-6) and by Strachey ( 1959: 83-6). Strachey points out that, in Freud's later work, it is only the form taken by anxiety that is to be understood as stemming from the experience of birth.
-379-
tie to his mother and to have led him to reflect further on the theory of anxiety -- an early example of the value of direct observation.
It was the publication of Rank Trauma of Birth in 1924, Freud relates in an addendum to Inhibitions, Symptoms and Anxiety ( 1926a), that 'obliged me to review the problem of anxiety once more'. In this work Rank had taken up the suggestion which, as we have seen, had first been thrown out by Freud, 'that the affect of anxiety is a consequence of the event of birth and a repetition of the situation then experienced . . . But', Freud continues, 'I could make no headway with his idea that birth is a trauma, states of anxiety a reaction of discharge to it and all subsequent affects of anxiety an attempt to "abreact" it more and more completely' ( SE 20: 161). Instead, what Freud does in his courageous re-examination of theory is to return to the safe ground of empirical observation -- which brings him back once more to separation anxiety.
284
In reading Inhibitions, Symptoms and Anxiety we find that Freud wrestled with the theoretical problems of anxiety through seven chapters, in the course of which he abandons a favourite hypothesis, namely that anxiety represents a direct transformation of libido. His reason for doing so lies in his recognition that, whereas formerly he had supposed anxiety to be the product of repression, an examination of clinical material suggests that, on the contrary, repression is a consequence of anxiety ( SE 20: 109). As a result of this, at the beginning of the eighth chapter he concludes ruefully: 'Up till now we have arrived at nothing but contradictory views about it [anxiety]. . . . I therefore propose to adopt a different procedure. I propose to assemble, quite impartially, all the facts that we do know about anxiety without expecting to arrive at a fresh synthesis' (p. 132 ). After a brief diversion he proceeds:
Only a few of the manifestations of anxiety in children are comprehensible to us, and we must confine our attention to them. They occur, for instance, when a child is alone, or in the dark, or when it finds itself with an unknown person instead of one to whom it is used -- such as its mother. These three instances can be reduced to a single condition, namely that of missing someone who is loved and longed for. But here, I think, we have the key to an understanding of anxiety . . . anxiety appears as a reaction to the felt loss of the object (pp. 136 -7).
-380-
Up to this point Freud is working from empirical data, data moreover which are now amply confirmed. Nevertheless he still remains puzzled, as others have also been, as to how to explain his observations. Why should there be this reaction of anxiety? It 'has all the appearance', he remarks, 'of being an expression of the child's feeling at its wits' end, as though in its still very undeveloped state it did not know how better to cope with its cathexis of longing' (p. 137 ). Today we can draw on a more sophisticated theory of instinctive behaviour to frame a hypothesis which regards the 'cathexis of longing' as the essence of the problem. Fifty years ago, however, such ideas on instinctive behaviour were unknown; instead, Freud was under the impression that the child's attachment could be understood only in terms of secondary drive and that the only primary needs are those of the body.
Freud therefore proceeds:
'The reason why the infant in arms wants to perceive the presence of its mother is only because it already knows by experience that she satisfies all its needs without delay. The situation then, which it regards as a 'danger' and against which it wants to be safeguarded is that of non-satisfaction, of a growing tension due to need, against which it is helpless. '
This, he continues, is 'analogous to the experience of being born . . . What both situations have in common is an economic disturbance caused by an accumulation of amount. 3 of stimulation which require to be disposed of. It is this factor, then, which is the real essence of the "danger" . . . ' and which he terms the 'traumatic situation'. To avoid this, an infant, by a process of learning, displaces 'the danger it fears . . . from the economic situation on to the condition which determined that situation, viz. the loss of object. It is the absence of the mother that is now the danger; and as soon as that danger arises the infant gives the signal of anxiety, before the dreaded economic situation has set in' (pp. 137 -8).
In considering Freud's every approach to the problem of anxiety it is necessary constantly to bear in mind that, from the earliest days of his psychoanalytic theorizing onwards, he adopts as his basic postulate that the nervous system has the function of getting rid of stimuli and that
285
the greatest catastrophe that can befall it is that of being overwhelmed by stimuli (see this volume, Chapter 5). Such theorizing constitutes what Freud describes as the economic viewpoint, and is cast sometimes in
-381-
terms of a psychical energy that builds up and is either discharged in action or else becomes dammed up, and sometimes in terms of excitation or stimulation that similarly varies in quantity. The 'dreaded economic situation' that Freud believes threatens an infant who is separated from his mother is none other than the damming up of psychical energy that cannot be discharged.
As a consequence of his re-examination of the problem, Freud concludes that anxiety has two sources. Anxiety from the first source arises as 'an automatic phenomenon', with physiological features that he believes may well be part of a response appropriate to the situation of birth. Such anxiety occurs whenever a traumatic situation 'is established in the id', that is in 'situation[s] of non-satisfaction in which the amounts of stimulation rise to an unpleasurable height without its being possible for them to be mastered psychically or discharged . . . ' ( SE 20: 137-41). Such traumatic situations are always characterized by helplessness. In Freud's formulation of this source of anxiety we see a direct descendant of his earliest theory, that advanced in his paper on the 'Anxiety Neurosis' ( 1895), in which he postulated that anxiety is developed when the nervous system is incapable of dealing with a mass of excitation.
Anxiety from the second source, Freud suggests, constitutes 'a rescue signal' designed to indicate that danger is impending. Since it requires foresight, such anxiety can 'only be felt by the ego' ( SE 20: 140). It is indeed the task of the ego so to imagine the danger situation in advance that it can restrict 'that distressing experience to a mere indication, a signal' (p. 162 ). Freud proceeds to list a number of danger situations, each corresponding to a particular developmental phase, which, if allowed to develop, would result in a traumatic situation: among these are birth, loss of object (namely mother), fear of father, and fear of superego (pp. 146 -7).
In his account of this second source of anxiety Freud lays much emphasis on the elements of foresight and expectation: 'The individual will have made an important advance in his capacity for self-preservation if he can foresee and expect a traumatic situation . . . which entails helplessness, instead of simply waiting for it to happen' (p. 166 ). Before anxiety can arise from this source, therefore, a fair degree of cognitive development is necessary.
Although, as has been indicated, Freud conceives of separation anxiety itself as no more than a signal and as being
-382-
developed through a process of learning, as indeed is necessary if it is based on foresight, it is evident that he is not entirely satisfied with that conclusion. At the end of the book ( SE 20: 168) he returns yet again to 'the puzzling phobias of early childhood' and hazards that perhaps, as in other species, the fear of loss of object may be a built-in response: thus he refers to an 'archaic heritage' and 'vestigial traces of the congenital preparedness to meet real dangers'. These reflections, and also the similar ones in regard to the child's tie that are found in the Outline ( 1940), which were noted in the appendix to the first volume of the present work,
286
suggest that towards the end of his life Freud was moving towards a formulation not very different from that advanced here.
It is far more important, however, that in this late work Freud finally clarifies what is held here to be the true relatedness of separation anxiety to mourning and defence. Previously, as he candidly admits, he had been confused. Not only had he supposed that repression is antecedent to anxiety, but he had also found it difficult to believe that anxiety as well as grief can be a response to loss of object. Now he sees that sequence clearly: anxiety is the reaction to the danger of loss of object, the pain of mourning is the reaction to the actual loss of object, and defences protect the ego against instinctual demands which threaten to overwhelm it and which can occur all too readily in the absence of the object ( SE 20: 164-72). This formula has not commonly been adopted by later theorists.
Ernest Jones
When Jones ( 1927) first advanced his theory of aphanisis it is evident, from the absence of references to Inhibitions, Symptoms and Anxiety, that he was unaware of Freud's latest train of thought. Furthermore, there is evidence also that he was still unaware of the importance of the attachment to mother (irrespective of the child's sex). It must therefore be noted that Jones's theory of aphanisis -- that 'the fundamental fear is [of] the total and of course permanent extinction of the capacity (including opportunity) for sexual enjoyment' -- was advanced without any reference to the present topic. The only mention of separation anxiety is in reference to weaning as a pregenital precursor of castration and to a girl's fear of separation from her father.
Two years later, however, Jones ( 1929) strives to integrate his own theory of aphanisis with Freud's theory of signal
-383-
anxiety. The union is uneasy and the resulting theory appreciably more complex than either taken singly. One of several difficulties is that Jones is still not aware of the tie to the mother irrespective of the child's sex. 1 Since the combined theory has been little called upon, to delineate it is unnecessary. As a broad generalization, it may be said that Jones accepts Freud's view regarding signal anxiety, believing it to be 'purposely provoked by the ego so as to warn the personality' of the possible approach of serious dangers, and then, in describing these serious dangers, adds to Freud's conception of what constitutes the 'traumatic situation' his own notion of aphanisis.
Melanie Klein
Whereas Jones developed his theory of anxiety independently of Freud's and later attempted to marry the two, Melanie Klein not only developed hers independently of Freud's but has frequently underlined the differences between them. Anxiety, in her judgement, is to be understood in terms of the death instinct, to which Freud never referred in this connection, and therefore in terms of aggression. Her views in regard to anxiety in general, which were taking shape between 1924 and 1934, and to separation anxiety in particular, are set out fully in her paper 'On the Theory of Anxiety and Guilt' ( 1948b, in Klein et al. 1952). They represent the only formulation made by a psychoanalyst which is both substantially different
287
from the formulations discussed by Freud and has had significant influence on theory and practice.
In Inhibitions, Symptoms and Anxiety, following a line of argument already advanced in Introductory Lectures ( SE 16: 407-8), Freud explicitly rejected the notion that fear of death is a primary anxiety and concluded instead that it is a later and learnt fear. 2 Melanie Klein differs: 'I do not share this view because my analytic observations show that there is in the unconscious a fear of annihilation of life. ' This she assumes must
____________________
1 See, for instance, his reference to the external danger arising from withdrawal of object,
'e. g. the mother in the boy's case' (p. 311, my italics).
2 It is interesting to note that Sylvia Anthony ( 1940) in her study of the genesis of children's
ideas of death reached a similar conclusion. Furthermore, she believes that it is through its equation with separation that death acquires its emotional significance: 'Death is equated with departure. . . . To the young child death means, in the departure context, its mother's death -- not its own. '
-384-
be the response to the death instinct: 'Thus in my view the danger arising from the inner working of the death instinct is the first cause of anxiety' ( Klein et al. 1952: 276). This, she suggests, is felt by an infant 'as an overwhelming attack, as persecution', and a persecution, moreover, which is first experienced at birth: 'We may assume that the struggle between life and death instincts already operates during birth and accentuates the persecutory anxiety aroused by this painful experience. ' From this argument she draws an important conclusion regarding the infant's first object relations: 'It would seem', she says, 'that this experience [i. e. birth] has the effect of making the external world, including the first external object, the mother's breast, appear hostile' ( 1952: 278). In another paper ( 1946) she summarizes her view in a sentence: 'I hold that anxiety arises from the operation of the death instinct within the organism, is felt as fear of annihilation (death) and takes the form of fear of persecution' ( Klein et al. 1952: 296). It is against this backcloth -- that anxiety is the result of the perpetual activity of the death instinct and that the newborn infant is already burdened with persecutory anxiety -- that Klein presents her views on separation anxiety.
Starting from Freud's distinction between objective anxiety (arising in connection with a known external danger) and neurotic anxiety (arising in connection with an unknown and internal one) ( Freud 1926a, SE 20: 165 and 167), Klein (1948b) sees both as contributing to the infant's fear of loss. She describes their nature as follows: objective anxiety arises from 'the child's complete dependence on the mother for the satisfaction of his needs and the relief of tension'; neurotic anxiety 'derives from the infant's apprehension that the loved mother has been destroyed by his sadistic impulses or is in danger of being destroyed, and this fear . . .
contributes to the infant's feeling that she will never return'. Were Klein to postulate that this depressive anxiety only developed in later infancy, she would not be diverging materially from Freud's view but only expanding it at an important point. This, however, is not her position. She emphasizes that in her view both sources of anxiety are present from the beginning and are constantly interacting. Because of this, 'no danger-situation arising from external sources could ever be experienced by the young child as a purely external and known danger' ( Klein et al. 1952: 288). On this her own statements and those of her colleagues are consistent. In discussing the cotton-reel incident Klein dissociates
288
-385-
herself explicitly from Freud's view and concludes, 'when [an infant] misses [his mother], and his needs are not satisfied her absence is felt to be the result of his destructive impulses' (pp. 269 -70). In the same volume it is claimed by Susan Isaacs that always 'mental pain has a content, a meaning, and implies phantasy. On the view presented here, "he behaves as if he were never going to see her again" 1 means his phantasy is that his mother has been destroyed by his own hate or greed and altogether lost' ( Klein et al. 1952: 87).
These passages seem to make it clear that in their explanations of separation anxiety Melanie Klein and her colleagues see depressive anxiety as virtually its sole component. This, however, is not so since elsewhere they emphasize that the relationship to the mother is itself 'a first measure of defence . . . The dependence on the mother and fear of loss of her, which Freud regards as the deepest source of anxiety, is from our point of view (the self- preservative) already a defence against a greater danger (that of helplessness against destruction within)' ( Joan Riviere in Klein et al. 1952: 46-7). 'From the very beginning,' she writes, 'the internal forces of the death instinct and of aggression are felt to be the cardinal danger threatening the organism' (p. 44 ). Since these forces are let loose during a separation experience, in the final analysis separation anxiety is seen as a response to the threat of destruction within. Clearly, this theory is very different from that of Freud and also from that advanced here. Whereas Freud gives primacy to anxiety that arises from 'an accumulation of amounts of stimulation' that he conceives as resulting from separation, Melanie Klein and her colleagues give primacy to persecutory anxiety.
It should, however, be added that in various passages Klein refers also to birth as constituting an anxiety-provoking trauma, and seems at times to subscribe to the birth-trauma theory of separation anxiety. Thus, following a passage already quoted above ( 1952: 296), she writes: 'Other important sources of primary anxiety are the trauma of birth (separation anxiety) and frustration of bodily needs. ' Nevertheless, although postulating these additional sources of anxiety, she quickly brings them within the ambit of persecutory anxiety by attributing to an infant a tendency always to suppose fear to be aroused by an object. After having earlier expressed the opinion that 'the fear of the destructive impulse seems to attach itself at once to
____________________
1 Isaacs quotation, from Inhibitions, Symptoms and Anxiety, is from the 1936 English
translation ( London: Hogarth, p. 167). -386-
an object', she completes her statement regarding the trauma of birth and the frustration of bodily needs thus: 'and these experiences too are from the beginning felt as being caused by objects. 1 Even if these objects are felt to be external, they become through introjection internal persecutors and thus reinforce the fear of the destructive impulse within' (p. 296 ).
In evaluating Melanie Klein's views it is essential to realize that her main theoretical outlook was formed in the years preceding the publication of Freud Inhibitions, Symptoms and Anxiety, and that, unlike Freud, who in the final formulation of his theory took anxiety arising from separation experiences as his point of departure, Klein had already developed her theory of anxiety before she gave any attention to separation from mother as a situation that provokes
289
anxiety. The first occasion she discusses it is in 1935 in her paper on 'The Psychogenesis of Manic-depressive States'.
When we look back on the early papers of Melanie Klein we remain impressed by her observation that anxiety and unconscious aggression often coexist, particularly when there is an unusually anxious and intense attachment of one person to another. In my judgement, however, she assumed too readily that aggression both precedes and causes anxiety so that, instead of recognizing conscious and unconscious aggression as a common response to separation and as constituting an important and frequent condition for the exacerbation of separation anxiety, she came to see aggression as the single source of anxiety; and furthermore, by identifying the child's tie to his mother with orality, was led into making implausible assumptions about the mental life of infants during their early months and thence into creating a theoretical superstructure that is far from convincing. This has had two unfortunate results. On the one hand, some of her critics have failed to appreciate the value of certain parts of her contribution; on the other, her followers have been slow to recognize that, significant though depressive and persecutory anxieties may sometimes be, the origin of separation anxiety cannot be understood in such terms, and, more important, that disturbances of the mother--child relationship that arise during the second and many subsequent years can have a far-reaching potential for pathological development.
____________________
1 Freud is not favourable to this type of theory. He writes: 'A child who is mistrustful in this
way and terrified of the aggressive instinct which dominates the world is a theoretical construction that has quite miscarried' ( SE 16: 407).
-387-
Anna Freud
Whereas Melanie Klein has written much about separation anxiety but has recounted few observations of how infants and young children actually behave in situations of separation, Anna Freud was one of the first to record such observations but until recent years has discussed their theoretical implications singularly little. As in the case of Klein, it looks as though a main reason was that her theoretical orientation was already set before Freud's fresh appraisal of the nature and genesis of anxiety appeared. Inhibitions, Symptoms and Anxiety is not referred to in her book The Psycho-analytical Treatment of Children ( 1946), which dates from 1926, 1927, and 1945; and, though a chapter is given to processes of defence in relation to the source of anxiety or to danger, there is no reference in The Ego and Mechanisms of Defence ( 1936) either to separation anxiety or to loss of object. Until her experiences with babies and young children in the Hampstead Nurseries during the war, Anna Freud seems to have given little attention to these problems.
In the two modest volumes published with Dorothy Burlingham ( Burlingham & Freud 1942; 1944), observation is sharp and description telling. Of children aged between one and three years they write: 'Reactions to parting at this time of life are particularly violent . . . This new ability to love finds itself deprived of the accustomed objects and his greed for affection remains unsatisfied. His longing for his mother becomes intolerable and throws him into states of despair' ( 1942:51). Yet, despite this clear understanding of the distress that is implicit in these responses, neither in these two volumes nor in the papers published by Anna
290
Freud during the subsequent decade are such manifestations related in any systematic way to anxiety in general or to separation anxiety in particular.
Instead, one has the impression that Burlingham & Freud were unprepared for the intensity of the responses they saw in the Nurseries and puzzled how to explain them. For instance, there is a passage ( 1942: 75-7) where they express the belief that perhaps if separations could be arranged more gradually all would be well: 'It is not so much the fact of separation to which the child reacts as the form in which the separation has taken place. ' In another passage (p. 57 ) the distress of a child between three years and five years seems to be attributed, entirely, to his belief that separation is a punishment--'To overcome this guilt he overstresses all the love which he has ever
-388-
felt for his parents'--a comment which suggests that in their view there would be no distress at this age were there not guilt and persecutory anxiety. Perhaps they get nearer the truth when in these same passages they refer to 'the natural pain of separation' and to the fact that 'unsatisfied longing produces in him a state of tension which is felt as shock'.
Whenever during that period Anna Freud broaches a theoretical interpretation of these responses or of the long-term results of separation (e. g. 1952; 1953), she takes for granted that the child's tie to his mother is to be accounted for by the theory of secondary drive. Since the infant has no needs but those of his body, his interest is at first confined to anyone who meets those needs; in so far as there is anxiety at separation from mother, it is a result of the fear that bodily needs will go unmet. Her views are perhaps most clearly expressed in an address to medical students ( 1953). After describing her conception of how attachment grows in the well-cared-for child she proceeds:
On the other hand, in cases where the mother has carried out her job as provider indifferently, or has allowed too many other people to substitute for herself, the transformation from greedy stomach-love to a truly constant love attachment is slow to come. The infant may remain too insecure and too worried about the fulfilment of his needs to have sufficient feeling to spare for the person or persons who provided for them [my italics].
This conclusion is a logical outcome of the secondary-drive theory of the child's tie and of Freud's version of the signalanxiety theory of separation anxiety.
More recently, in a book published in 1965, Anna Freud describes several 'forms' taken by anxiety during the early years, each of which she believes to be characteristic of a particular phase in the development of object relations. The sequence of forms runs as follows: 'archaic fears of annihilation, . . . separation anxiety, castration anxiety, fear of loss of love, guilt . . . ' Separation anxiety (and also fear of annihilation) is held to be characteristic of, and confined to, the first phase in the development of object relations; that is described as a phase of 'biological unity between the mother--infant couple, with the mother's narcissism extending to the child, and the child including the mother in his internal "narcissistic milieu" . . . '. During subsequent phases, forms of anxiety other than separation anxiety are thought to occur. For example, the third
-389-
291
phase, described as that of object constancy, is believed to be characterized by fear of the loss of the object's love. ?
Contributions by Other Exponents of Ego Psychology
The theories advanced by Anna Freud during earlier years are subscribed to also by Nunberg ( 1932), Fenichel ( 1945), and Schur ( 1953; 1958). In his two carefully reasoned studies of anxiety, Schur makes the commonly held assumption that in man the biologically given components of behaviour are strictly limited. In the later paper, in which he draws extensively on ethological data and concepts, he details what he believes they comprise. On the one hand, he postulates the presence of fight and flight reactions characteristic of the phase of development which begins with the ability to perceive external objects. On the other, he postulates an earlier phase ('the undifferentiated phase') during which 'all danger is, due to the infantile development specific for man, "economical", inner danger', namely danger arising from an accumulation of excitation that springs from unmet bodily needs. It is from this, specifically human, source of anxiety that he regards separation anxiety as developing as a learnt derivative: 'The realization that an external object can initiate or end a traumatic situation displaces the danger from the economic situation to the condition which determines that situation. Then it is no longer hunger that constitutes danger for the child but it is the absence of the mother. ' Although he discusses various dangers which he thinks 'may be based on innate givens', nowhere does he consider the possibility that loss of mother may be one of them.
After he has come to recognize the importance of separation anxiety Kris ( 1950) makes a serious effort to incorporate it in his theorizing. But his views are based more on inference from previous theory than on a reassessment of the data; in particular he is concerned, like Schur, to cast them in a form compatible with Hartmann's ego psychology. This leads him to place great emphasis on a distinction between the danger of losing the love object and the danger of losing the object's love. Although this distinction was referred to briefly by Freud ( 1926a), the way that Kris elaborates it is his own. On theoretical grounds he postulates that the danger of losing the love object is concerned solely with anaclitic (namely bodily) needs and is not concerned with a particular love object. Conversely, the development of a 'relationship to a permanent personalized love object that can no longer easily be replaced' he postulates
-390-
to occur synchronously with the development of responsiveness to the danger of losing the object's love; it represents, in his opinion, 'a decisive step in ego development'.
This hypothetical association is not, however, borne out by observation. Anxiety reactions to the loss of a particular love object are to be seen some months before it is reasonable to credit a human infant with awareness of the danger of losing the object's love and before the twelve- month age-limit suggested by Kris ( 1950). As is emphasized in the previous volume (Chapter 15), the responses mediating attachment behaviour both in man and in lower species tend quickly to focus on a particular figure; and there can be no reason to suppose that their doing so represents an important step in ego development. In the event, therefore, the theoretical distinction advanced by Kris must be regarded as mistaken.
292
The crucial connection between anxiety as the reaction to the danger of losing the object and the pain of mourning as the reaction to its actual loss, which Freud arrives at in the final pages of Inhibitions, Symptoms and Anxiety, has been little recognized. Only in the work of Melanie Klein and Therese Benedek is it given much place. Helene Deutsch ( 1937) explicitly divorces the two: anxiety is an infantile response, she holds, grief and mourning more mature ones. 'The early infantile anxiety', she writes, 'we know as the small child's reaction to separation from the protecting and loving person. ' When the child is older, on the other hand, 'suffering and grief [are] to be expected in place of anxiety' (p. 14, my italics). Moreover, separation anxiety in the older individual is to be understood always as a regression to infancy, and occurs in situations where 'grief . . . threaten[s] the integrity of the ego, or, in other words, if the ego [is] too weak to undertake . . . mourning' (p. 14 ). This differentiation by maturity does not stand examination, however. In the responses of infants and young children to loss of mother, elements of grief are undoubtedly present. Conversely, as Therese Benedek among others has recorded, anxiety is the rule even in adults when they are separated for any length of time from someone they love.
For many years Therese Benedek has been concerned with problems of separation from, and reunion with, loved persons, and with responses to loss and bereavement; and as a result of her clinical work she has had a lively awareness of the farreaching significance of separation anxiety and of its close relatedness to anxiety and mourning. In describing responses to
-391-
separations, reunions, and bereavements occurring during wartime, she frequently speaks of separation as a trauma in itself, and she generalizes boldly: 'The universal response to separation is anxiety' ( Benedek 1946: 146). She also recognizes that the experience of being separated, or the expectation of being separated, from a loved person leads to a sharp increase in longing for his company. In a later paper ( 1956) she notes that a crying fit in an infant is by no means always caused 'by a commanding physiologic need such as hunger and pain, but by the thwarting of an attempt at emotional (psychologic) communication and satisfaction'.
All these observations can be parsimoniously explained in terms of the theories regarding attachment, separation anxiety, grief and mourning that are advanced in the present work. Nevertheless, although her original training took place in Budapest (see Appendix to Volume I), Benedek does not accept these simpler hypotheses. Instead, in all her theorizing she is committed to a secondary-drive theory of the child's tie to his mother with all its complications and disadvantages. Thus the increase in longing evident in adults at separation, which can hardly be considered other than a natural and normal response, is explained as due to a regression to oral dependency. Indeed, as in so much theorizing deriving from the concept of dependence, Therese Benedek tends at times to theorize as though all attachments to loved persons were undesirable regressions to an infantile state.
Nowhere in Benedek's writings is there any systematic discussion of separation anxiety; but, in the later paper ( 1956) referred to above, two separate theories appear to be adumbrated. The first is similar to Freud's signal-anxiety theory; the second is concerned with the danger of ego disorganization.
Still struggling with the same problem that Freud was wrestling with thirty years earlier, she asks why an infant should respond to 'the frustration of a "dependent" wish' by crying. Reverting to the belief that crying is related intrinsically only to the experiences of hunger and
293
pain, she concludes that 'he responds to the lack of participation from the adult as to a complete interruption of the symbiosis, as if he were abandoned and hungry' (p. 402, my italics).
Since, however, she is not altogether confident that crying is to be understood as anxiety, and believes that anxiety proper is a response to the danger of ego distintegration, Benedek advances another view. This is that the young child has to turn
-392-
to his mother to preserve his ego integration when faced with the 'anxiety, humiliation and shame of failure'. In the case of the older child, 'his ego can maintain itself by its own resources' (pp. 408 -9). Thus, although the clinical data she presents are consistent with the theory advanced here, Benedek's interpretations remain firmly embedded within the traditional paradigm.
In much of her theorizing, especially in her use of the concept of symbiosis, Margaret Mahler ( 1968) follows Therese Benedek; and in attributing a distinct form of anxiety to each phase of the development of object relations she follows Anna Freud. Nevertheless, despite the similarity of their postulated phases, the phase of development to which Mahler attributes separation anxiety is not the same as the one to which Anna Freud attributes it. Whereas Anna Freud regards separation anxiety as a response specific to 'infringements of the biological mother-infant tie' during the first phase of development, Mahler holds that separation anxiety is attributable properly only to a later phase, namely the phase 'after the beginning of object constancy has been achieved'. This she puts in the third and fourth years. The form of anxiety that Mahler attributes to the first phase of development, the symbiotic phase, is a fear of self- annihilation, the reasoning being that at that phase 'loss of the symbiotic object' is thought to amount 'to loss of an integral part of the ego itself'. This mode of theorizing is close to that of Spitz.
Although, like most other analysts, Spitz is an adherent of the secondary-drive theory to account for the child's tie to his mother and endorses Freud's version of the signal-anxiety theory of separation anxiety ( 1950), he advances, in addition, a variant of that theory. This is a theory of 'narcissistic trauma'. After outlining his views on the development of object relations from a phase of narcissism (first three months) through a phase of pre-objectal relations (second three months) to a phase of true object relations (third three months), he proceeds:
It is in the third quarter that true objects appear for the first time. They now have a face, but they still retain their function of a constituent part of the child's recently established Ego. The loss of the object is therefore a diminution of the Ego at this age and is as severe a narcissistic trauma as a loss of a large part of the body. The reaction to it is just as severe.
From other passages, in which he insists on the warning function
-393-
of anxiety and its dependence on learning and foresight, it is clear that in Spitz's view anxiety is a signal to warn against the danger of a narcissistic trauma. This is a fresh variant of the
294
signal-anxiety theory: the traumatic situation to be avoided is, this time, one in which narcissism is threatened.
It should be noted that much of Spitz's theorizing about anxiety turns around his concern to explain the anxiety exhibited by an infant of seven or eight months when confronted by a stranger, which he was the first to term eight-months' anxiety; the anxiety exhibited at separation from a loved object is less in his mind. In view of his empirical work this may seem surprising, until we realize that his observations of deprived infants were not concerned with the immediate responses to separation, namely protest, distress, and anxiety, or with responses after reunion, but were largely concentrated on responses seen during the later phases of separation, namely grief and depression. As a result he had no opportunity to observe the continuum of response from separation anxiety to grief and mourning.
The approach of Sandler & Joffe to these problems follows fairly closely the approaches of Kris and Spitz. The traditional theory of secondary drive is adopted to account for the child's tie to his mother, together with the concept of dependency. In keeping with their basic model, moreover, they place almost exclusive emphasis on the feeling states produced in a child by the presence or absence of his mother and make little attempt to relate these feeling states either to instinctive behaviour or to the survival value of mother's presence and the increased risk attendant on her absence. Thus they describe 'the role of the object' in a child's life as 'that of a vehicle for the attainment of the ideal state of well-being'. Conversely, 'loss of the object signifies the loss of an aspect of the self', i. e. that 'part of the self-presentation which . . . reflects the relation to the object' ( Joffe & Sandler 1965, their italics).
For Sandler & Joffe, therefore, as for Freud, the situation at all costs to be avoided is not so much the actual loss as the traumatic overwhelming of the ego to which loss leads. In terms of the Sandler & Joffe model, the traumatic situation to be avoided is described as a 'disruption of the individual's feeling state' ( Sandler & Joffe 1969).
Summing up the differences they see between their theoretical interpretation and my own, Joffe & Sandler ( 1965) conclude:
-394-
Object loss may bring about acute mental pain through creating a wound in the self. This view coincides with what Abraham and others have described as the 'severe injury to infantile narcissism' which object loss entails. And although Bowlby has maintained ( 1960b) that such a statement misses the true significance of object loss, we take the view that it contains its essence.
Such contrasting positions are, of course, a simple consequence of our having adopted different paradigms.
Other Contributors
In view of Sullivan's insistence that psychiatry is the study of interpersonal relationships, it is not unexpected that he sees all anxiety as a function of the child's relationship to his mother and other significant people. Nevertheless his position is different from that advanced here, especially in the primacy that he gives to the role of learning; for he regards anxiety as being exclusively a product of the mother's attitude. When mother is approving, her child is content;
295
when she is disapproving, her child is anxious. Despite his great emphasis on 'need for contact' and 'need for tenderness' and the strong terms in which he refers to the experience of loneliness -- 'really intimidating' and 'terrible' ( 1953: 261) -- that separation from a loved object can of itself induce anxiety appears to be explicitly ruled out. Thus, in a final chapter, he indicates features that man has in common with other species, namely bodily needs and 'even our recurrent need for contact with others'. These he contrasts with features 'restricted to man and some of the creatures he has domesticated', which include 'the experience of anxiety' (p. 370 ). His assumption that anxiety is confined to domesticated species follows from his assumption that it results from processes of training and learning: 'there is nothing I can conceive in the way of interpersonal action about which one could not be trained to be anxious' (pp. 370 -1). Even 'the experience of intense anxiety' which gives rise to repression is conceived as resulting from ill-conceived educational methods (p. 163 ). 1
Although in Sullivan's view the induction of anxiety remains
____________________
1 Dr Mabel Blake Cohen has emphasized that Sullivan did not regard such 'training' as a
product only of conscious parental attitudes: ' Sullivan recognized that unconscious attitudes or tensions in the parents' interactions with the child were of considerably more importance than conscious planned behaviour' (personal communication).
-395-
something of a mystery -- 'the character of situations which provoke anxiety is never completely to be grasped' (p. 190 ) -- it is nevertheless evident that in effect he sees it as always connected with processes of child-training. Since he believes that a main anxiety- inducing sanction used by a mother is restriction or denial of tenderness (p. 162 ), he comes at times near to my own concept of separation anxiety and its exacerbation by threats of abandonment. What seems to escape Sullivan, however, is that distress and anxiety can be and often are direct consequences of lack of tenderness and of separation per se; and that threats to restrict tenderness would be ineffective were that not so. While aware that loneliness can be a devastating experience for adolescents and adults. Sullivan seems unaware that it is even more distressing for infants and young children; indeed, there are passages in which he seems specifically to exclude that that is so: 'Loneliness, as an experience which has been so terrible that it practically baffles clear recall, is a phenomenon ordinarily encountered only in preadolescence and afterwards' (p. 261, my italics).
Reading Sullivan's work one gets the impression that he had never observed young children and that he was only partially aware of the close attachment they form to particular people and of the sense of security that mere proximity to a loved figure brings. The 'need for contact with others, often felt as loneliness', is identified, not with need for a genital or a parent-child relationship, but with gregariousness in animals (p. 370 ); his conviction that 'no action of the infant is consistently and frequently associated with the relief of anxiety' (p. 42 ), which overlooks the relief an infant commonly exhibits when clutching his mother, is a main plank in his theorizing. Because of this, he seems never to have grasped the reality of separation anxiety and, therefore, despite his close attention to the problems to which it gives rise, it remains almost impossible to attribute to him any particular theory of its nature and origin. It is probably for the same reasons that neither grief nor mourning plays any significant part in his system of psychopathology.
296
In the theorizing of Phyllis Greenacre ( 1952) separation anxiety, and grief and mourning seem also to be omitted. Instead, experiences during the birth process and the first weeks of postnatal life are advanced as major variables to account for a later differential liability to neurosis (see Chapter 16 of this volume).
Rank's views regarding birth trauma have already been
-396-
referred to. In his early papers Fairbairn, who sees separation anxiety as the mainspring of all psychopathology, follows Rank closely in regard to its origins: Fairbairn's postulate ( 1943) that birth anxiety is 'the prototype of all the separation anxiety which is subsequently experienced' is the counterpart of his postulate that a return-to-womb craving accounts for the child's tie. It should be added, however, that these views are peripheral to Fairbairn's main theoretical position ( Fairbairn 1952), which is in all other respects consistent with the theory of frustrated attachment advanced here. In a late paper ( 1963) in which he gives a synopsis of his views he writes: 'The earliest and original form of anxiety, as experienced by the child, is separation anxiety. '
Others have also founded their psychopathology on the central role of separation anxiety and some have adopted a frustrated attachment theory to account for it. For instance, as long ago as 1935, Suttie, holding the view that the child's attachment to his mother is the result of a primary 'need for company', saw anxiety as 'an expression of apprehension of discomfort at the frustration, or threatened frustration, of this all-important motive'. A year later Hermann ( 1936) expressed an almost identical view. He relates anxiety to the urge to seek and cling to mother: 'Anxiety is basically the feeling of being left on one's own in the face of danger. Its expression is a seeking for help and at the same time a seeking for mother.
