No More Learning

The most important component in learning not to fear situations formerly feared, Bandura ( 1968) finds, is that the observer should see that the feared situation can be approached and dealt with without there being any bad consequences. The identity of the person observed (model) and the degree to which the observer can identify with him are found to be of much less significance. Even watching a sequence on film can have a reassuring effect, provided always that the consequences of the model's actions are clearly depicted.
The process of learning that something is harmless from direct observation of the experience of others is very different, it should be noted, from merely being informed by another person that a situation is harmless. All those who have made a systematic study of the problem report that simple explanation and reassurance have only very limited effect, a finding that will come as no surprise to clinicians.
Fortunately, in the ordinary course of events, a child growing up in a family has endless opportunities to learn from observation that many of the situations that make him afraid are in
151
fact harmless. Parents, older brothers and sisters, neighbours and schoolfellows are continually and without knowing it providing a child with this indispensable information.
Observational Learning combined with Guided Participation
This method requires much more from the model than giving the subject opportunity for simple observational learning. It is evident none the less that every sensible parent is constantly providing it. The method consists in the model's first demonstrating in action that the feared situation holds no danger and then encouraging the other person -- child or adult -- to tackle the situation himself. Once again it appears that the crucial part of the process is that the learner should discover, this time for himself, that approaching and tackling the situation can be done without untoward consequences. The efficacy of the method was commented on by several of the early students of
-193-
children's fear behaviour (e. g. Jones 1924a; Jersild & Holmes 1935a), and their findings have been amply confirmed by Bandura and his colleagues in a number of recent experiments. In one experiment, reported by Bandura ( 1968), a study was made of a group of adolescents and adults who suffered acute fear of snakes. The subjects were divided into four subgroups and given four different sorts of treatment:
the now standard desensitization procedure of imagining increasingly alarming situations with snakes and at the same time engaging in deep relaxation exercises;
observing a graduated film depicting young children, adolescents, and adults engaging in progressively more fearprovoking interactions with a large harmless snake;
observing the therapist engage in a carefully graduated series of such procedures and at each step being aided by the therapist to engage in the same procedures, so that gradually the subject is himself led first to touch and stroke the snake, then to grasp the snake round the middle while the therapist holds its head and tail, and so on step by step until the subject is able to allow the snake in the room with him, to retrieve it, and finally to let it crawl freely over him; only when a subject has accomplished one step without fear is he encouraged to go on to the next;
receiving no treatment but, like subjects in the other subgroups, being tested for fear of snakes both at the start of the experiment and at the end of it, thus providing a control group.
When subjects in the four subgroups were tested at the end of their treatment by being required to engage in increasingly daring activities with snakes, those who had both observed the therapist interact with the snake and themselves taken part in the graduated exercises with it showed much the least fear. Subjects in subgroups (a) and (b) were less fearful than before but had not benefited as much as had those in subgroup (c). Finally, those in the control group showed as much fear of snakes at the end of the proceedings as they had done at the beginning.
In commenting on his results Bandura suggests that the striking efficacy (for these subjects) of observational learning combined with guided participation rests on two features of the method: first, the subject's fear is reduced sufficiently to enable
-194-
152
him to start a process of interaction with the feared object; and second, after he begins to interact, he discovers for himself that it has no disagreeable consequences. Bandura lays emphasis on the point that to be successful the method has to be carefully graduated so that at no stage is fear of more than modest intensity aroused.
In the context of this work, perhaps the most important aspect of Bandura's findings is the key role played in his technique by a trusted and encouraging companion. Not only does the therapist perform the fear-arousing acts, but he stands by while the subject tries the same measures himself, encouraging him at every success and reassuring him after any failure. Only in the presence of such a companion is a subject likely to feel confident enough to tackle the problem in active fashion and so to discover for himself what the consequences really are.
A second valuable lesson from the work of behaviour therapists is that it is essential to work forward in small steps so that the fear aroused is never beyond low intensity. Once fear at high intensity is aroused, it is found, the subject may well be back where he began. It is of interest that the careers of men who later become astronauts appear to be built in a similar way, moving steadily from one modest success to another in unbroken series ( Korchin & Ruff 1964). These findings are referred to again in Chapter 21.
It is fortunate that most parents seem to know intuitively that no good comes from allowing a child to become acutely frightened. They also know that what allays fear more certainly than anything else is their own presence. As the Newsons write of their sample of four-year-olds and their mothers:
Two out of three of all our children have definite and recurrent fears of which the mother is aware. Once she realizes that the child is frightened, she will go through a series of remedies until she finds one that works: and that a remedy is effective is the main consideration to most mothers, even if it does upset the household, for few are unsympathetic to fear. There are no certain methods, and some fears are immune to endless ingenious expedients: the parents can only hope that the child will eventually 'grow out of it'. In general, mothers tend to favour a mixture of explanation and simple cuddling; and these usually at least have a soothing effect, even if they do not always drive the fear away ( Newson & Newson 1968).
-195-
Experiences and processes that increase susceptibility to fear
It is argued in Chapter 6 that 'it is no less natural to feel afraid when lines of communication with base are in jeopardy than when something occurs in front of us that alarms us and leads us to retreat'. As a consequence, an individual's increased tendency to respond to situations with fear can be a result of either (or both) of two distinct types of experience. One is an experience in a particular situation that has led the person henceforward to become especially prone to avoid or withdraw from that situation. The other is uncertainty about the availability of his attachment figure (s). As a rule a specially alarming experience is likely to lead to an increased susceptibility to respond with fear in that specific situation only; whereas uncertainty about the availability of attachment figures results in increased susceptibility to respond with fear to such a wide range of situations that the person concerned is often referred to as suffering from 'free-floating anxiety'.
153
Since the remaining chapters of this volume are concerned with susceptibility to anxiety about the availability of attachment figures, here we deal mainly with experiences that increase a person's susceptibility to be afraid of specific situations.
Frightening Experiences
Jersild and his colleagues and also the Newsons present evidence that in very many cases when an individual exhibits unusually intense fear of a particular situation the origin can be traced to a specific experience connected with that situation.
In describing their four-year-olds the Newsons remark that, when a child's previous experiences are known, his fear is often seen to be 'reasonable', even though it may now seem exaggerated. As examples, they describe: a child who had intense horror of mud which dated from a summer holiday during which her feet were trapped in wet sand so that, when the other children ran off, she was unable to follow; a child who would not go near water after she had fallen into a river; and a child terrified of anyone in a white coat after he had been shouted at and held down while being x-rayed ( Newson & Newson 1968).
Evidence of a similar sort and from two separate sources is reported by Jersild & Holmes: (a) from parents about factors that may have contributed to a child's having developed
-196-
unusually intense fear of some particular situation ( Jersild & Holmes 1935b), and (b) from young adults about what factors they believe have been responsible for their having themselves developed intense and/or persistent fear of some situation ( Jersild & Holmes 1935a). For obvious reasons neither source is adequate and a great deal of further research is required.
Like the Newsons, Jersild & Holmes describe a number of cases in which a child's fear of a specific situation is reported by a parent to have developed in a thoroughly intelligible way. Examples are a child frightened of all objects resembling a balloon, whether on earth or in the air, following an operation during which a gas balloon had been used for an anaesthetic; and another child afraid of a familiar pet canary after having been frightened by the sudden hooting of an owl in the zoo. All such cases can be understood as due to a child's generalizing from too small a sample.
Similarly, the group of young adults report that in many instances fear of a particular situation had followed an alarming experience they had had as children. Examples include witnessing an accident, returning home to find the house had been burgled, witnessing an explosion, and mother being ill.
Since not all children become persistently afraid after a particularly alarming experience, specific conditions are presumably responsible. Of possible candidates, compound situations of which one component is being alone seem especially likely. It is perhaps noteworthy that in none of the examples quoted above is it stated whether the child was alone or with a trusted companion. In future studies of what appear retrospectively to have been traumatic situations, therefore, exact details of all the conditions obtaining are necessary.
154
There is, of course, a large literature regarding experiences that have led individual animals to become persistently afraid of specific situations ( Hebb 1949). Animals, however, cannot be made afraid by stories heard or by threats uttered, as humans can.
Stories Heard
A major cause of persistent and/or intense fear was said by the young adults questioned by Jersild & Holmes ( 1935a) to have been hearing lurid tales, some true and some fictional. Other evidence suggests that this may be a more frequent cause of certain individuals coming to fear certain situations than is
-197-
often supposed. An example given by Jersild & Holmes ( 1935b) is of an unprecedented number of young children reported to be afraid of wolves during the period when the song 'Who's afraid of the big bad wolf? ' was popular. In view of the difficulties a child has in distinguishing fact from fiction and in making realistic assessment of potential danger, already touched upon in Chapter 10, this finding should not surprise us. It seems likely that fear arising from such misunderstandings, though intense enough at the time, usually becomes modified once the individual's grasp of the world improves.
Situations of several sorts that are feared by some children and adults and not by others can be understood as culturally determined. For example, several studies report a difference of incidence in regard to fear of certain situations dependent on socio-economic class. In interviews of 400 children aged between five and twelve years in the vicinity of New York City a higher proportion of children from public schools than from private schools reported fear of robbers and kidnappers and also of supernatural happenings ( Jersild & Holmes 1935a). In their study of 482 children aged from six to twelve years in Buffalo, New York, based on interview data from mothers, Lapouse & Monk ( 1959) report a higher incidence of fear of wars, floods, hurricanes and murders, of fire and of being kidnapped among whites of lower socio-economic class than among upper-class whites. A difference in the same direction is reported by Croake ( 1969) who interviewed 213 children between the ages of eight and twelve years in South Dakota and Nebraska.
Many other differences in incidence between groups reported in the literature seem likely to be due to cultural influences.
Threats
In answering the questionnaire administered by Jersild & Holmes ( 1935a) many of the young adults were unable to give any clear account of how or why they had developed intense and/or persistent fear of some situation. Nevertheless, in examining the reasons that were given, the researchers were struck by how large a part deliberate threats of horrifying consequences seemed to have played in a number of cases. Some of those threats had been made by older children, sometimes perhaps to tease but at other times with serious intent. Other threats had been made by parents, or occasionally a schoolteacher, as a means of discipline. Some of these threats were of physical punishment. More often they were an exploitation of a
-198-
155
child's tendency to fear one of the natural clues, notably darkness, isolation, or abandonment.
Unfortunately Jersild & Holmes found that it was not possible to make an exact count of 'apparently deliberate attempts to frighten' but they record some of the more extreme instances. It is a disturbing list. For example, if the answers to the questionnaire are to be believed, a child's fear of the dark had been exploited either by his having been punished by being locked in a dark room or cellar, or by his having been threatened that that would be done. In a few cases a child's fear of the dark had been amplified by its being alleged that the dark room was filled by such things as vicious rats or dreadful monsters.
Another type of threat used for disciplinary purposes, reported both by Jersild & Holmes ( 1935a) and by the Newsons ( 1968), is one entailing separation from parents. The threat can take one of several forms. A child can be threatened that he will be sent away, or that some alarming figure will come to take him away, or that his mother will go away and leave him. There is reason to believe that many children are exposed to threats of this kind, and also that such threats play a far larger part in increasing a person's susceptibility to separation anxiety than has yet been realized by psychiatrists. Evidence for these statements is given in later chapters (15, 18, 19), and some of the reasons why the role of these threats has been so seriously underestimated are discussed in Chapter 20.
The Key Role of Experience
In clinical circles great emphasis is often placed on the existence of cases in which a much raised susceptibility to respond with fear in a situation cannot be accounted for, apparently, by any experience of the kind so far discussed. Resort is then had to more complex explanations, often turning on fear of 'internal dangers'. The position taken here is that such explanations are invoked far too readily. In some cases highly relevant experiences are unknown to the patient or his relatives; in others they are known about but for one of many reasons are deliberately not reported. In yet other cases, experiences are known about but go unreported because they are thought not to be relevant or because the clinician appears uninterested or unsympathetic. In other cases again experiences are mentioned by the clinician hardly registers them because he is guided by theories that give them no place. Finally, it is not uncommon for fear that is
-199-
aroused by one situation to be attributed erroneously either by patient or by clinician to another.
A major theme of this volume is that no fear-arousing situation is missed or camouflaged as often as is fear that an attachment figure will be inaccessible or unresponsive.
-200-
156
Chapter 14
Susceptibility to Fear and the Availability of Attachment Figures
Throughout all this ordeal his root horror had been isolation, and there are no words to express the abyss between isolation and having one ally.
G. K. CHESTERTON, The Man who was Thursday
Forecasting the availability of an attachment figure
Enough has been said about the conditions that arouse fear to make plain how crucial a variable it is to be with or without a trusted companion. In the presence of a trusted companion fear of situations of every kind diminishes; when, by contrast, one is alone, fear of situations of every kind is magnified. Since in the lives of all of us our most trusted companions are our attachment figures, it follows that the degree to which each of us is susceptible to fear turns in great part on whether our attachment figures are present or absent.
But man does not live entirely in the present. As a child's cognitive capacities increase he becomes capable of foreseeing the possible occurrence of many sorts of situation, including those that he knows would arouse fear. And, of the many feararousing situations that a child, or older person, can foresee, none is likely to be more frightening than the possibility that an attachment figure will be absent or, in more general terms, unavailable when wanted.
It has already been noted (Chapter 1) that, in reference to an attachment figure, presence is to be understood as implying ready accessibility rather than actual and immediate presence, and absence implies inaccessibility. Still further amplification is required, however, since accessibility in itself is not enough. Not only must an attachment figure be accessible but he, or she, must be willing to respond in an appropriate way; in regard to someone who is afraid this means willingness to act as comforter and protector. Only when an attachment figure is both accessible and potentially responsive can he, or she, be said to be truly
-201-
available. In what follows, therefore, the word 'available' is to be understood as implying that an attachment figure is both accessible and responsive.
In this chapter three distinct propositions are introduced, each of which is basic to the thesis of this work. The first is that, when an individual is confident that an attachment figure will be available to him whenever he desires it, that person will be much less prone to either intense or chronic fear than will an individual who for any reason has no such confidence. The second proposition concerns the sensitive period during which such confidence develops. It postulates that confidence in the availability of attachment figures, or a lack of it, is built up slowly during the years of immaturity -- infancy, childhood, and adolescence -- and that whatever expectations are developed during those years tend to persist relatively unchanged throughout the rest of life. The third proposition concerns the role of actual experience. It
157
postulates that the varied expectations of the accessibility and responsiveness of attachment figures that different individuals develop during the years of immaturity are tolerably accurate reflections of the experiences those individuals have actually had.
Each of these three propositions is controversial, though each is much more controversial in some quarters than in others.
To psychoanalysts who adopt an object-relations theory of personality the first proposition is familiar enough: in terms of that theory, a person's confidence, or lack of it, in the availability of an attachment figure is regarded as resulting from his having either introjected, or failed to introject, a good object. By contrast, to those unfamiliar with object-relations theory, or perhaps alternatively with ethology, the proposition may be novel and even surprising.
The second proposition lies somewhere between, on the one hand, a view that attributes a high degree of plasticity to personality structure even during mature years and, on the other, a view, stemming especially from the work of Melanie Klein, that regards plasticity of personality as diminishing rapidly after the earliest months of infancy and as falling to a low ebb as soon as the first year or two are past. The view adopted here contrasts with both. It holds that the period during which attachment behaviour is most readily activated, namely from about six months to about five years, is also the most sensitive in regard to the development of expectations of the availability of attachment figures; but that nevertheless sensitivity in this
-202-
regard persists during the decade after the fifth birthday, albeit in steadily diminishing degree as the years of childhood pass.
The third proposition, which concerns the role of experience, may seem fairly self-evident to many but has nevertheless been extremely controversial in psychoanalytic circles. It is a crucial question, since on how it is answered turn not only preventive measures but also therapeutic techniques. The controversy is referred to repeatedly in this and later chapters.
These three propositions, each of which can in principle be tested, provide the scaffolding for the rest of the volume. It is claimed that each is inherently plausible, that none is contradicted by evidence of weight, and that together they enable the evidence available to be fitted together in a way that makes sense.
Working models of attachment figures and of self
The states of mind with which we are concerned can conveniently be described in terms of representational or working models. In the first volume it is suggested that it is plausible to suppose that each individual builds working models of the world and of himself in it, with the aid of which he perceives events, forecasts the future, and constructs his plans. In the working model of the world that anyone builds, a key feature is his notion of who his attachment figures are, where they may be found, and how they may be expected to respond. Similarly, in the working model of the self that anyone builds a key feature is his notion of how acceptable or unacceptable he himself is in the eyes of his attachment figures. On the structure of these complementary models are based that person's forecasts of how accessible and responsive his attachment figures are likely to be should he turn to them for support. And, in terms of the
158
theory now advanced, it is on the structure of those models that depends, also, whether he feels confident that his attachment figures are in general readily available or whether he is more or less afraid that they will not be available -- occasionally, frequently, or most of the time.
Intimately linked to the type of forecast a person makes of the probable availability of his attachment figures, moreover, is his susceptibility to respond with fear whenever he meets any potentially alarming situation during the ordinary course of his life.
The theory proposed can be formulated in two steps: from the
-203-
early months onwards and throughout life the actual presence or absence of an attachment figure is a major variable that determines whether a person is or is not alarmed by any potentially alarming situation; from about the same age, and again onwards throughout life, a second major variable is a person's confidence, or lack of confidence, that an attachment figure not actually present will none the less be available, namely accessible and responsive, should he for any reason desire this. The younger the individual the more influential is the first variable, actual presence or absence; up to about the third birthday it is the dominant variable. After the third birthday forecasts of availability or unavailability become of increasing importance, and after puberty are likely in their turn to become the dominant variable.
Although the concepts of working models and forecasts derived from working models may be unfamiliar, the formulation adopted is no more than a way of describing, in terms compatible with systems theory, ideas traditionally described in such terms as 'introjection of an object' (good or bad) and 'self-image'. The advantages claimed for the present concepts are that they allow for greater precision of description and provide a framework that lends itself more readily to the planning and execution of empirical research.
Reflection shows that working models of attachment figures and of self can vary along many dimensions. One is that of simplicity versus sophistication (see Volume I, Chapter 17). Another is that of validity, which is discussed briefly later in this volume (Chapter 20). Yet another is the extent to which the roles of attachment figures, on the one hand, and of self, on the other, are differentiated. Let us consider the last.
Confidence that an attachment figure is, apart from being accessible, likely to be responsive can be seen to turn on at least two variables: (a) whether or not the attachment figure is judged to be the sort of person who in general responds to calls for support and protection; (b) whether or not the self is judged to be the sort of person towards whom anyone, and the attachment figure in particular, is likely to respond in a helpful way. Logically these variables are independent. In practice they are apt to be confounded. As a result, the model of the attachment figure and the model of the self are likely to develop so as to be complementary and mutually confirming. Thus an unwanted child is likely not only to feel unwanted by his parents but to believe that he is essentially unwantable, namely unwanted by
-204-
159
anyone. Conversely, a much-loved child may grow up to be not only confident of his parents' affection but confident that everyone else will find him lovable too. Though logically indefensible, these crude over-generalizations are none the less the rule. Once adopted, moreover, and woven into the fabric of the working models, they are apt henceforward never to be seriously questioned.
Whereas common sense might suggest that a person would operate with only single models of each of his attachment figures and of himself, psychoanalysts from Freud onwards have presented a great deal of evidence that can best be explained by supposing that it is not uncommon for an individual to operate, simultaneously, with two (or more) working models of his attachment figure(s) and two (or more) working models of himself. When multiple models of a single figure are operative they are likely to differ in regard to their origin, their dominance, and the extent to which the subject is aware of them. In a person suffering from emotional disturbance it is common to find that the model that has greatest influence on his perceptions and forecasts, and therefore on his feeling and behaviour, is one that developed during his early years and is constructed on fairly primitive lines, but that the person himself may be relatively, or completely, unaware of; while, simultaneously, there is operating in him a second, and perhaps radically incompatible, model, that developed later, that is much more sophisticated, that the person is more nearly aware of and that he may mistakenly suppose to be dominant.
How and why multiple models originate and persist raises difficult questions of defensive processes, to which attention is given in the third volume. The hypothesis of multiple models, one of which is highly influential but relatively or completely unconscious, is no more than a version, in different terms, of Freud's hypothesis of a dynamic unconscious.
In terms of the present theory much of the work of treating an emotionally disturbed person can be regarded as consisting, first, of detecting the existence of influential models of which the patient may be partially or completely unaware, and, second, of inviting the patient to examine the models disclosed and to consider whether they continue to be valid. In pursuing this strategy an analyst finds that how the patient perceives him (the analyst), and what forecasts the patient makes about his likely behaviour, are particularly valuable in revealing the nature of the working models that exert a dominant influence
-205-
in the patient's life. Because certain of these perceptions and forecasts appear to the analyst so clearly to be based on a patient's preconceptions about him and to be derived from working models that stem from experiences with other people during earlier years, rather than from current experience, how the patient perceives and conceives the analyst is often known as 'transference'. When an analyst interprets the transference situation he is, among other things, calling the patient's attention to the nature and influence of those models and, by implication, inviting him to scrutinize their current validity and, perhaps also, to revise them.
Seen in the perspective of Piaget's theorizing, the concept of transference implies, first, that the analyst in his caretaking relationship to the patient is being assimilated to some preexisting (and perhaps unconscious) model that the patient has of how any caretaker might be expected to relate to him, and, second, that the patient's pre-existing model of caretakers has not yet been accommodated -- namely, is not yet modified -- to take account of how the analyst has actually behaved and still is behaving in relation to him.
160
Some analysts argue that only those features of the model that are inappropriate to the current situation should be referred to as transference. In practice, however, it is often very difficult to disentangle what parts of a complex model are mistakenly applied to the analyst and what parts are in some degree truly applicable. As a result a tradition has grown up of referring to all aspects of a patient's conception of and attitude towards the analyst as transference. There is perhaps no harm in this, provided that the issue of which parts of the model are inapplicable to the analyst and which parts are in some degree applicable is kept constantly in mind.
Not infrequently, a striking feature of a patient's forecasts is his strong expectation of being abandoned by his analyst, an expectation by no means always fully conscious. During weekends and holidays, and especially during unexpected separations due to illness or other contingency, the way in which a patient behaves and the thoughts and feelings he expresses may be intelligible only on the hypothesis that he forecasts that the analyst will not return, and often also on the assumption that the analyst no longer wishes to see him. Not infrequently these forecasts, either conscious and expressed as fear or not conscious and expressed in some distorted form, persist in spite of assurances that they are mistaken. What is far more impor-
-206-
tant, moreover, they persist often in spite of repeated falsification in real life. 1
At the same time as he is drawing the patient's attention to the nature of the forecasts he (the patient) appears to be making, the analyst is striving, jointly with the patient, to understand how the models on which the forecasts are based may have come into being. During those inquiries it is often found that a model, currently active but at best of doubtful current validity, becomes reasonably or even completely intelligible when the actual experiences that the patient has had in his day-to-day dealings with attachment figures during all his years of immaturity are known. This leads again to the controversial question of the extent to which actual experience is of influence in the development of working models of self and others.
The role of experience in determining working models
There was a time when psychoanalysts were as loath as a Kraepelinian psychiatrist might be to attribute a patient's unfavourable models of attachment figures to his actual experiences. To make such attributions was deemed to be nai? ve, to be underrating the role of projection, and to be failing to give due weight to the contribution that the patient himself had made, and was making, to the misfortunes he experienced. Nowadays, thanks to the influence of Fairbairn, Winnicott, and others, fewer psychoanalysts take that view, which, it is held, is tenable only so long as a clinician limits himself to treating isolated patients, usually adults, and is uninterested in considering systematically what their day-to-day experiences have actually been: to what extent and in what ways has a patient had his attachment behaviour met, not only during early infancy, about which information is usually uncertain and often entirely speculative, but throughout all the later years of childhood also?
Probably no one who has worked for long in a family clinic, in which disturbed children and their parents are treated, still holds the traditional view that actual experience is of little consequence. On the contrary, in the course of that work it is found repeatedly that, when information about a child's experiences of interaction with parents and parent figures is
____________________
161
1 Although responses of these kinds are referred to in numerous case reports, I am unaware of any systematic empirical record of how one (or more) patient(s) responded to the separations, planned or unplanned, that occurred during the course of his (or their) analysis.
-207-
obtained -- usually partly from first-hand observation of members of the family who are interviewed together and partly from the family's history as it is pieced together, often only slowly and from a diversity of sources -- the forecasts the child makes of how attachment figures are likely to behave towards him are not unreasonable extrapolations from his experiences of the way in which they have behaved towards him in the past, and may perhaps still be behaving towards him in the present. Thus, whatever contributions to variations of personality are made by genetic biases and physical traumata, the contribution of family environment is certainly substantial.
From the viewpoint of the position adopted, adult personality is seen as a product of an individual's interactions with key figures during all his years of immaturity, especially of his interactions with attachment figures. Thus an individual who has been fortunate in having grown up in an ordinary good home with ordinarily affectionate parents has always known people from whom he can seek support, comfort, and protection, and where they are to be found. So deeply established are his expectations and so repeatedly have they been confirmed that, as an adult, he finds it difficult to imagine any other kind of world. This gives him an almost unconscious assurance that, whenever and wherever he might be in difficulty, there are always trustworthy figures available who will come to his aid. He will therefore approach the world with confidence and, when faced with potentially alarming situations, is likely to tackle them effectively or to seek help in doing so.
Others, who have grown up in other circumstances, may have been much less fortunate. For some the very existence of caretaking and supportive figures is unknown; for others the whereabouts of such figures have been constantly uncertain. For many more the likelihood that a caretaking figure would respond in a supportive and protective way has been at best hazardous and at worst nil. When such people become adults it is hardly surprising that they have no confidence that a caretaking figure will ever be truly available and dependable. Through their eyes the world is seen as comfortless and unpredictable; and they respond either by shrinking from it or by doing battle with it.
Between the groups of people with extremes of either good or bad experience lie groups of people with an almost infinite range of intermediate sorts of experience, who grow up to have expectations of the world to match. For example, some may have learnt that an attachment figure responds in a comforting
-208-
way only when coaxed to do so. They grow up supposing that all such figures have to be coaxed. Others may have learnt during childhood that the wished-for response can be expected only if certain rules are kept. Provided the rules have been moderate and sanctions mild and predictable, a person can still come confidently to believe that support will always be available when needed. But when rules have been very strict and difficult to keep, and
162
when sanctions on breaking them have been severe and especially when they have included threats to withdraw support, confidence is likely to wilt.
Sanctions of a damaging kind, which are used by many parents, include both refusing to respond to a child's approaches, for example by sulking, and threatening to leave the home or to send the child away. When used repeatedly, or even only occasionally but with intensity, such sanctions or threats of sanctions can have calamitous effects on a developing personality. In particular, because they deliberately cast grave doubt on whether an attachment figure will be available when needed, such threats can greatly increase a person's fear that he will be abandoned, and thereby greatly increase also his susceptibility to respond to other situations fearfully.
Admittedly the influence that these sorts of experience have on the development of personality and especially on susceptibility to fear and anxiety is still controversial. Evidence in support of the position adopted, introduced already in Chapter 16 of the first volume, is presented more fully in the coming chapters. It is hoped that those who adopt a different position, for example that experiences of the kinds described play at most a subordinate role in accounting for variations in personality development, will be stimulated to present the evidence on which they base their views. Only in this way can progress be made.
A note on use of the terms 'mature' and 'immature'
In many clinical circles the practice has developed of referring to personalities as 'mature' or 'immature'.
A person who approaches the world with confidence yet who, when in difficulty, is disposed to turn to trusted figures for support is often said to be mature. In contrast, both someone who is chronically anxious and permanently in need of support and someone who never trusts anyone are said to be immature.
The theory underlying this use of immature is that adult personality structures so described are held to be a consequence
-209-
of arrested development and to have remained in a state that, though normal for early childhood, is passed through during the course of healthy growing up and left well behind.
The theory advanced here and discussed further in the final chapter is different. It disputes that mental states either of chronic anxiety or of persistent distrust are characteristic of normal or healthy stages of development. Instead, it holds that the main cause of such deviations is that, during childhood, an individual's attachment behaviour was responded to in an inadequate or inappropriate way, with the result that throughout later life he bases his forecasts about attachment figures on the premise that they are unlikely to be available.
The resemblance of certain of these types of personality to the personalities typical of young children, especially in so far as individuals of both sorts require and often demand the constant presence and support of attachment figures, is held to be superficial only. In the case of a young child he has no means by which to make forecasts except over short spans of time. In the case of an 'immature' personality he not only has the means to make forecasts but the forecasts he makes, and makes with conviction, are that attachment figures will be unavailable unless he maintains constant watchfulness or is constantly humouring them.
163
Thus the common usage of the terms mature and immature is held to be inaccurate and misleading. A particularly adverse effect of using immature in this way is that it can, on occasion, lead a clinician to take a humouring patronizing attitude to the persons concerned, instead of recognizing that their behaviour is a legitimate product of bitter experience.
-210-
Chapter 15
Anxious Attachment and Some Conditions that Promote it
If there are quarrels between the parents or if their marriage is unhappy, the ground will be prepared in their children for the severest predisposition to a disturbance of sexual development or to a neurotic illness.
SIGMUND FREUD ( 1905b)
'Overdependency' or anxious attachment
In the opening pages of this volume vignettes are given (quoted from Burlingham & Freud 1944) of children, aged between two and four years and resident in the Hampstead Nurseries, who showed intensely possessive behaviour towards one or another nurse. Jim, for example, who had been in the nursery since the age of seventeen months, is described as having formed 'strong attachments' first to one young nurse and then to another who had successively looked after him. Towards each he was intensely clinging and possessive, and he refused to be left by them for a minute. Numerous other observers, including my colleagues Robertson and Heinicke, have also noted this type of behaviour whenever small children in a nursery setting are given opportunity to make an attachment to a member of staff, and the same behaviour is shown towards mother after they return home.
Clinging behaviour, either literal or figurative, can be seen at every age, during childhood, during adolescence, and during adult years. It goes by many names. Among adjectives used to describe it are 'jealous', 'possessive', 'greedy', 'immature', 'overdependent', and 'strong' or 'intense' attachment. For scientific and clinical purposes, it is argued, each word has drawbacks: because it derives from and implies obsolescent theory, or because it is ambiguous, or, and perhaps most important of all, because it carries with it an adverse value judgement that is held to be inappropriate and unhelpful.
Both 'jealous' and 'possessive', though accurate, are apt to be -211-
pejorative. The same is true of 'greedy', which is used mainly by those whose thinking is still influenced by the assumption that attachment derives from food and being fed.
'Strong' attachment and also 'intense' attachment are ambiguous: both of them, and the former especially, might be thought to imply a satisfactory state of affairs.
'Immature' derives from a theory of regression which, as indicated at the end of the previous chapter, is held to be out of keeping with the evidence.
164
Some of the ambiguities and false values concealed in the terms 'dependency' and 'overdependency' are emphasized in the first volume (Chapter 12). 1 Let us look further at their deficiencies and also propose an alternative term.
Perhaps no terms are used more frequently in the clinical literature than 'dependent' and 'overdependent'. A child who tends to be clinging, an adolescent reluctant to leave home, a wife or husband who maintains close contact with mother, an invalid who demands company, all these and others are likely sooner or later to be described with one of these words. Always in their use there is an aura of disapproval, of disparagement. Let us consider more closely the behaviour to which these terms are applied and how we are to evaluate the persons who come to be described by them.
Viewed in the perspective of this work, most persons described by clinicians as dependent or overdependent are ones who exhibit attachment behaviour more frequently and more urgently than the clinician thinks proper. Inherent in the terms, therefore, are the norms and values of the observer using them. This leads to many difficulties. One is that norms and values differ greatly not only between individuals but from culture to culture and from subculture to subculture. To take a crude example, behaviour that in some parts of the East might pass unnoticed, or even be encouraged, might in the West be condemned as childishly dependent. Another difficulty is that, even within a single culture, no useful evaluation of the behaviour can be made without knowledge of the conditions, organismic and environmental, in which it is shown. Ignorance of a child's age, of whether he is well or ill, of whether or not a person has recently experienced a shock, can play havoc with an observer's judgements. Individuals who are notoriously apt to be wrongly
____________________
1 For a discussion of how the concepts of dependency and attachment are related to one
another see Ainsworth ( 1972); overlap in their meanings is not complete. -212-
judged as overdependent are children who look older than they are, who are tired or unwell, or who have recently acquired a new sibling, and adults recently bereaved. Another example is a young woman during pregnancy or while caring for small children. In all such instances attachment behaviour is likely to be shown more frequently and/or more urgently than would otherwise be the case. In other words, in the conditions obtaining the behaviour may be well within normal limits and no adverse conclusions on the personality development of the individuals concerned would be appropriate.
There are, however, persons of all ages who are prone to show unusually frequent and urgent attachment behaviour and who do so both persistently and without there being, apparently, any current conditions to account for it. When this propensity is present beyond a certain degree it is usually regarded as neurotic.
When we come to know a person of this sort it soon becomes evident that he has no confidence that his attachment figures will be accessible and responsive to him when he wants them to be and that he has adopted a strategy of remaining in close proximity to them in order so far as possible to ensure that they will be available. To describe this as overdependency obscures the issue. Even the term 'separation anxiety' is not ideal. A better way to describe the condition is to term it 'anxious attachment' or 'insecure attachment'. This makes it clear that
165
the heart of the condition is apprehension lest attachment figures be inaccessible and/or unresponsive. For these reasons, therefore, and especially because it can be expected to enlist our sympathy, anxious attachment is the term to be used. It respects the person's natural desire for a close relationship with an attachment figure, and recognizes that he is apprehensive lest the relationship be ended.
The thesis of the present work is that, even though other causal factors may play some part in the development of this condition, those about which by far the most evidence is at present available are experiences that shake a person's confidence that his attachment figures will be available to him when desired. Alternative theories, some long entrenched, are considered in the next chapter.
The following descriptions by two working-class mothers of occasions when their young children went through a phase of 'overdependency' reveal the condition in what is believed to be its true light. The descriptions are taken from the survey of 700
-213-
four-year-old children in Nottingham undertaken by Newson & Newson ( 1968). Asked whether her daughter sometimes wanted to be cuddled, a miner's wife replied:
Ever since I left her that time I had to go into hospital (two periods, 17 days each, child aged 2 years), she doesn't trust me any more. I can't go anywhere -- over to the neighbours or in the shops -- I've always got to take her. She wouldn't leave me. She went down to the school gates at dinner time today. She ran like mad home. She said, 'Oh, Mum, I thought you was gone! ' She can't forget it. She's still round me all the time.
When asked the same question a lorry-driver's wife, whose husband had deserted three months earlier, replied:
Yes, all the time just lately -- only since he left. (What do you do? ) Well, if I'm not busy I sit down and nurse her, because -- you know -- she's continually clinging round me, she keeps saying, 'Do you love me? You won't leave me, Mummy, will you? ' -- and so I sit down and try to talk to her about it, you know; but I mean, at her age [about four], really you can't explain. And she used to dress herself; but since my husband's been gone, she's relied on me for -- well, every mortal thing I've had to do for her. At the moment I'm more or less letting her do what she wants. I mean, she's been upset in one way, and I don't want to upset her again. Because I did put her in a nursery just after he went, because I thought it might take her mind off things, you see, but anyhow the matron asked if I'd mind taking her away, because she said she just sat and cried all day long. I think she'd got it into her head that because her Daddy's gone, and me taking her there and leaving her all day, she p'raps thought I'd left her too, you see. So she was only there a fortnight, and then I took her away. But she's afraid of being left on her own, I mean, if I go to the toilet, I have to take her with me, she won't even stay in the room then on her own. She's frightened of being left.
In summing up their findings on children who exhibited overdependency and fear of separation the Newsons write: 'Most of these children's separation fears are reality-based, in that they or their mothers have been hospitalized or some other separation has already taken place. ' Nevertheless, there were
166
-214-
some children who had had such an experience without apparently having been left prone to separation anxiety, and other children who were prone to such fears without having had such an experience. Important though experience of actual separation is, therefore, it is clear that variables of other kinds contribute as well.
Among the most influential of these other variables, it seems likely, are, first, threats to abandon a child, made for disciplinary purposes, and, second, a child's recognition that parental quarrels carry with them the risk that a parent may depart. In the light of present evidence, it seems extremely probable that, as Suttie ( 1935) and Fairbairn ( 1941) long ago suggested, threats to abandon a child are the most influential of all. None the less, it must not be forgotten that those threats have the tremendous power they do have only because for a young child separation is itself such a distressing and frightening experience, or prospect.
For that reason, therefore, we return once more to our point of departure, the effects on a young child of being separated from his mother figure.
In the two sections that follow we consider, first, children who are being reared in a residential setting without any permanent mother figure and, second, children who are being reared mainly at home with mother but who, for various reasons and for shorter or longer periods, have been separated from her.
Anxious attachment of children reared without a permanent mother figure
The most systematic data yet available on the attachment and fear behaviour of children reared without a permanent mother figure are provided by Tizard & Tizard ( 1971). They compare the social and cognitive development of two-year-old children cared for in residential nurseries in England with that of children brought up in ordinary families.
In recent years there have been great changes in the organization of residential nurseries in Britain. Not only are links with a child's family encouraged but, within the nursery itself, attempts are made to provide living conditions that are nearer than in the the past to ordinary family living. Apart from babies under twelve months, who are cared for in a separate unit, children live in groups of six, of varying ages up to about five or six years, and are cared for in 'private' accommodation by their own nurse and her assistant. In addition, in some nurseries a system
-215-
obtains whereby each nurse is encouraged to give special attention to one or two children, as a rule from a group other than the one in which she usually works; she takes the child out in her free time, buys him little presents, sometimes puts him to bed, or takes him to her home for weekends.
Although this type of re? gime is a big advance on some of the impersonal re? gimes of the past, examination shows that, so far as mothering care goes, it still falls very far short of what obtains in an ordinary working-class home in present-day London.
167
For their study Tizard & Tizard selected fifteen boys and fifteen girls (ten white and five coloured in each case), aged two years, who had been healthy full-term babies, whose health had throughout been good, and who had entered a nursery before the age of four months and had remained there since. All but one were illegitimate Half were visited by mother, who still hoped to be able to care for them; the others had been offered for adoption, but for various reasons adoption had been delayed.
A contrast group, similarly composed by age, sex, and health record, but confined to white English children, was selected from among working-class children living in their own homes and in intact families. For reasons of research convenience, any child whose mother was working full time and who had an older sibling of less than school age was excluded.
The aim of the study was to compare children in the two groups in respect of both their cognitive and their social development. Various cognitive tests were therefore administered, 1 and opportunity was taken to observe the children's responses first to the advent of a stranger and then to the brief departure of the caretaker from the room. In addition, to obtain further information regarding attachment behaviour, a child's caretaker was asked a series of detailed questions about it and also certain particulars regarding the child's experiences with potential attachment figures and with other persons. Both research workers engaged in the project were women.
In reporting results it is useful to begin with particulars of the opportunities that children in each of the groups had to make attachments. Comparison shows there were great differences.
For the thirty two-year-olds living in their families, mother
____________________
1 Results of the cognitive tests showed the mean of the nursery children to be two months
below the norm and three months below the mean of the family children. The inferiority of the nursery children was due mainly to their failures on the verbal subtests ( Tizardet al. 1972).
-216-
was principal attachment figure for twenty, father for four, and both parents equally for five. In one case, where father was away from home, it was a maternal uncle. The total number of figures towards whom attachment was shown was strictly limited, with an average of four per child in the group. Following an attachment figure around the house was reported as a regular activity for all but four of the children.
In contrast to this picture of focused attachment, the attachment behaviour of the nursery children was directed diffusely. By most of them the behaviour was shown in some degree towards a large, indeed an indeterminate, number of people, including usually 'anyone he knows well'. Despite this, however, each of the children had a preference. Provided a child's natural mother visited once a week or more, she was said always to be the preferred figure: 'the child would be very excited to see her and distressed when she left'. Similarly, when a child had a 'special nurse' who took him out (and no mother who visited regularly) the special nurse was always the child's preferred figure, despite the fact that she would probably see him for only a few minutes on most days. Thus, whereas the family children had almost constant contact with their preferred attachment figures, the nursery children saw very little of theirs.
168
Furthermore, the nursery children were not permitted to follow anyone out of the room in which they were.
Despite every endeavour by the nursery authorities to stabilize the children's social relationships the results were found to be disappointing. Since entering the nursery twenty months earlier, a majority of the nursery children had had at least twenty different people caring for them for periods of a week or longer, in comparison with an average of two in all for the family children. Even during the course of a single week the average number of nurses looking after the nursery children was found to be six. Furthermore, in the lives of nursery children members of staff were found to come and go at irregular intervals, often disappearing for days or weeks at a time, and sometimes never returning.
When data on attachment and fear behaviour for children in the two groups were compared the nursery children proved to be significantly more anxious 1 in their attachments and significantly more afraid of a stranger.
____________________
1 In their report Tizard & Tizard refer to the 'intensity' of the nursery children's attachment
behaviour as being 'much greater' than that shown by the family children. -217-
Information given by the nurses showed that the attachment behaviour of the nursery children both to the preferred figure and to the rest of the staff was much more anxious than that of the family children (as reported by mothers). For example, twenty-four of the thirty nursery children were apt to cry when their favourite caretaker left the room compared with thirteen of the thirty family children; those who cried regularly numbered ten and two respectively. When the caretaker returned, all but two of the nursery children would run to be picked up (and most of them did so usually), in comparison with only four of the family children (none of them usually). Thus, whereas about two-thirds of the family children took mother's coming and going from the room as a matter of course, all but a handful of the nursery children were upset when an attachment figure left and wanted to be picked up by her when she returned.
For children in each group a standard procedure was used to measure fear of strangers. Every child was assessed in his own living-room with his own caretaker (nurse or mother) present. During the first five minutes the child sat on his caretaker's lap, while the research worker chatted to the caretaker. Then the research worker made a series of standardized overtures to the child: she started with a greeting; as a second step she invited him to come over to look at a picture book; and lastly she invited him to sit on her lap. The child's response to each move was rated on a seven-point scale.
When, during the second phase, the research worker invited the child to come over, only fifteen of the thirty nursery children did so in comparison with twenty-six of the family children. When the nursery children were invited to sit on her lap only eight accepted, while six of them cried and ran away. Of the family children sixteen accepted the invitation and none ran away.
After the researcher had made her initial overtures to a child she engaged the caretaker in a few minutes' further conversation and then asked her to leave the room briefly and to leave the door ajar. The child's response to this was rated on a four-point scale. Towards the end of
169
the session the caretaker was again asked to leave the room and the child's response was again rated. When left alone with the research worker on the first occasion, six of the nursery children ran out of the room; and even on the second occasion later in the session five of them were still unwilling to stay with her. On both occasions
-218-
every one of the family children was ready to stay with the researcher.
Finally, at the end of the session, the research worker again invited the child to sit on her lap. The nursery children, although most of them were bolder by now, remained much more cautious than the family children. Two of the nursery children still ran away and cried; and none of them smiled and chatted as did eleven of the family children as they climbed up onto the research worker's lap.
Whereas all the findings cited so far from this study support our hypothesis, there are one or two others that do not. In particular, fearfulness, for example of dogs, was reported by nurses and mothers respectively 'with equal frequency' for children of the two groups. Considering the very different responses of the children to the strange research worker, we are perhaps entitled to question the validity of this information, which comes from informants whose standards may well not be comparable.
In this context a case report by Schnurmann ( 1949) is of interest. This describes how a little girl of two and a half years, who was being cared for in the Hampstead Nurseries, developed fear both of going to bed and of dogs. Although the symptoms are described as phobic and the writer seeks to explain them in terms of castration anxiety resulting from the child's observation of sex differences, it is apparent from the record that the onset of symptoms was closely related in time to the cessation of her mother's daily evening visits, and the remission of symptoms to the resumption of visits. The relation of phobic symptoms to anxious attachment is considered at length in Chapters 18 and 19.
In view of the very different experience with attachment figures that children in a residential nursery have compared with children in a family setting, it is hardly surprising that the attachment behaviour of the nursery children should be so much more anxious or that they should be more prone to fear strangers. Nor would it be surprising were there to be a substantial difference between the working models of attachment figures built by children in the two groups and used by them as the basis of their future predictions about the accessibility and responsiveness of such figures. For, whereas a family child lives in a stable and predictable world with accessible and usually responsive attachment figures, a nursery child, even in a modern nursery, lives in a highly unpredictable world in which his
-219-
preferred attachment figure is usually inaccessible, while subsidiary ones come and go almost at random.
Anxious attachment after a period of separation or of daily substitute care
After a period of separation, especially when it has been spent with strange people, it is usual for small children who are being brought up in families to be more anxious and clinging than
170
they were before. This finding, which constitutes part of the basic data on which the present thesis rests and which is well illustrated by the two examples (quoted from Newson & Newson 1968) given in the first section of this chapter, has perhaps ceased to be a matter of controversy. What is debatable concerns the factors that lead one child to recover confidence and another not to. ?
Effects of a Brief Stay in Hospital
A study by Fagin ( 1966) of the behaviour of two groups of children after return home from a short stay in hospital (varying from one to seven days) casts some light on this problem. The thirty children of one group had had mother staying with them in hospital, the thirty of the other group were there without her, though they were visited daily. The children were matched for age, between eighteen and forty-eight months, but not for sex. Interviews held with the mothers prior to the hospitalization suggested that attitudes towards child-rearing did not differ between mothers in the two groups, nor did the mothers differ in regard to their desire to remain with their child in hospital.
For each child, behaviour at one week and again at one month after his return home, as reported by mother at interview, was compared with what it had been before he entered hospital, as reported by her at that time.
Both at one week and at one month following return home the behaviour of the unaccompanied children was reported as markedly disturbed in comparison with what it had been prior to hospital. Significant differences occur in all the usual fields. In particular, these children were reported as being much more upset by a temporary brief separation than they were before hospital and as more 'dependent'. By contrast, children who had been accompanied in hospital by mother showed none of these adverse changes. Indeed, mothers' reports showed them to have developed favourably in all these respects, which suggests that, as MacCarthy and his colleagues ( 1962) have found, mother's
-220-
presence on a potentially upsetting occasion had given a child additional confidence that she would always be available in emergency. 1
These findings are in keeping with those of similar studies. What is of particular interest in Fagin's study is the differential effect on children in the two groups of having a mother who was rated by the interviewer as being very 'irritable'. Of the accompanied children who had an irritable mother none showed any adverse effects of their stay in hospital; whereas, of the unaccompanied children, those who had an irritable mother were even more affected by their experience than were those whose mother was rated as more equable.
A number of other studies point firmly to the view that a separation has a specially adverse effect on children whose parents are inclined to be hostile or to threaten them with separation as a disciplinary measure, or whose family life is unstable. Indeed, evidence is accruing fast that parent-child relations, both before and after the event, play a very large part in accounting for the differential outcome of a separation.
Effects of Periods of Residential Substitute Care
171
Moore ( 1964; 1969a and b) has reported a number of findings of much interest to our thesis. In a longitudinal study of London children, starting at birth with 223 cases and reducing to 167 at the age of six years, Moore was able to investigate the shortterm effects on children of separations and other discontinuities of care, and also differences in the behaviour, at the age of six or seven years, of children who, during their early years, had had different types of experience. Information regarding behaviour was obtained: (a) from interviews with mothers (two before the child was born, five during the child's first year, four
____________________
1 Limitations of Fagin's study are that data were derived entirely from mothers' reports and
that, in regard to ailment and length of stay, the two groups were not matched. Twenty-one of the accompanied children were in hospital for hernia operation or tonsillectomy and were out again within two days. Only nine of the unaccompanied children had a similar experience; thirteen had had respiratory or alimentary infections and had been away for between three and five days. It is therefore possible that these differences of experience might account for some of the reported differences in subsequent behaviour, though the findings of MacCarthy, Lindsay & Morris ( 1962) do not support that interpretation.
-221-
during the next two years, and annually thereafter); (b) from psychological tests and observations of the children carried out regularly at the Study Centre; and (c) from interviews with the heads and staffs of any day nurseries or nursery schools that the children attended. Most of the results are expressed in terms of differences between the average frequencies of behaviour in children in a series of samples differing in regard to previous experience.
A number of the children had spent periods of a week or two with relatives, by way of a holiday for the child or for the parents or both. The commonest behaviour of children under three on reunion with mother was clinging around her, sometimes after initial reserve. Although such clinging often passed off after two or three days, in 30 per cent of cases it persisted for a matter of weeks. Moore ( 1969b) concludes: 'It is clear that for most young children separation from the mother is in itself a stressful experience', and that they are especially vulnerable during their second and third years.
Moore reports much evidence that shows that whether disturbances persist or fade turns in great measure on the stability of the home and the attitudes of the parents. Details are given of three of his comparisons.
(a) There were six children who had experienced one or more stays in a residential nursery between the ages of nine and thirty months. All but one, who was backward, were upset on return home, 'showing it in aggressive behaviour, head banging, fear of strangers and/or increased dependence on their mothers'. Four other children had had a similar experience before the age of nine months; of these, two were similarly upset. When all these ten children had reached the age of eight years, two were found to be 'reasonably well-adjusted' and eight were not. It was noted that, whereas the two who were reasonably well adjusted came from homes in which family relationships were good, the eight who presented difficulties, mainly of an aggressive uncontrolled sort, came from homes which were breaking up (two cases) or where relationships were otherwise disturbed.
172
(b) Fifteen children who came from stable homes had experienced episodic separations during their first four years. Separations had been in various settings and of varying lengths. In many cases a child had stayed with relatives; in others he had been in a hospital or nursery. When the lengths of time away were summed for each child, summed lengths ranged from
-222-
five to twenty-three weeks. Not only did these children come. from stable families but, apart from the separations, they had been cared for throughout by mother. At the age of six years their behaviour could be compared with that of two similar groups of children:
those who had been cared for by mother throughout without having experienced any separations;
those who had not only experienced a variety of separations but whose home life had been generally unsettled, including frequent changes of daily caretaker during the pre- school years.
Judged by mothers' reports, the children from stable homes who had experienced these episodic separations were rather less given to attention-seeking than were those who had not been separated. This suggests that, for children cared for as a rule by mother in a stable family, episodic separations of the length and kinds described, although in many cases reported to have been very distressing at the time, do not leave obviously adverse effects on the children's later attachment behaviour. Nevertheless in drawing conclusions it is necessary to be cautious since we do not know how such children would react to a new fear-arousing situation.