By contrast, a baby is quieted by being rocked or patted and by
engaging
in non-nutritive sucking (see Volume I, Chapter 14).
Bowlby - Separation
Like every word denoting emotion, fear has as referent both how we suppose a person to be feeling
-87-
and how we predict he is likely to behave (see Volume I, Chapter 7). Since there is reason to think that fear behaviour has hitherto been given far too little attention, that is where we start. Fear Behaviour
Let us examine the various forms of behaviour that are commonly held to be indicative of fear. They include, of course, the initial forms of behaviour, such as posture, expression, and
76
incipient action, that lead us to infer that a person or animal is feeling afraid, and also the less subtle and more active forms of behaviour that often, but not always, follow. Both in ordinary life and in systematic field observation there is a large array of distinct forms of behaviour that it is common practice to group together as being indicative of fear. They include wary watching combined with inhibition of action, a frightened facial expression accompanied perhaps by trembling or crying, cowering, hiding, running away, and also seeking contact with someone and perhaps clinging to him or her. When we ask why all these diverse forms of behaviour should be grouped together, we find the following four reasons:
a. many of these forms of behaviour, though not all, tend to occur either simultaneously or sequentially;
b. events that elicit one of these forms tend to elicit others also (though not necessarily all the others);
c. most of them seem plainly to serve a single biological function, namely protection;
d. when asked how they are feeling, persons behaving in these ways commonly describe
themselves as feeling afraid or anxious or alarmed.
Although these are good reasons for grouping such diverse forms of behaviour together, there are nevertheless risks in doing so. In particular, the conditions that elicit one form of fear behaviour may differ in certain respects from those that elicit another form; and the autonomic responses that accompany one form may well differ from those that accompany another. In animals the distinctiveness of the forms is attested by experiment. Hinde ( 1970) discusses work by Hogan that suggests that, at least in young animals, not only may freezing and withdrawal be separate systems of behaviour, elicited by different types of external stimulation, but they may even be mutually inhibiting. Further evidence on this issue is given in Chapter 8.
-88-
A point to be especially noted, and a corner-stone of the present argument, is that, in ordinary usage, grouped under this single heading of behaviour indicative of fear are to be found forms of behaviour that have at least three distinct kinds of predictable outcome: (a) immobility, (b) increased distance from one type of object, and (c) increased proximity to another type of object. The contrast between the last two outcomes is especially important. For, on the one hand, is behaviour that increases distance from persons and objects that are treated as though they were threatening; on the other is behaviour that reduces distance from persons and objects that are treated as though they provided protection. Naturally, both these types of behaviour do not always occur. Yet they occur together with sufficient frequency to enable us to take the combination for granted. When we flush a rabbit we expect it not only to run from us but to run to cover. When a child is afraid of a barking dog we expect him not only to withdraw from the dog but to retreat towards a parent figure.
Now the usual practice of including under a single heading, that of behaviour indicative of fear, forms of behaviour that have such different predictable outcomes is of great significance. Yet it can very easily make for confusion. In particular it has often tempted psychologists, for example McDougall ( 1923), and also others, to postulate a single all-embracing 'instinct of fear'. An alternative theory and one that keeps far closer to observed data is that we are dealing, not with some single comprehensive form of behaviour, but with a heterogeneous collection of interrelated forms, each elicited by a slightly different set of causal conditions and each having a distinctive outcome. In the sense defined in Part II of the first volume, each such form can be regarded as an example of instinctive behaviour.
77
In order to sort out these different forms of behaviour a first step is to examine how attachment behaviour and fear behaviour are related to one another.
Withdrawal behaviour and attachment behaviour
It may already have been noticed that, of the three forms of behaviour with such different predictable outcomes that are habitually treated as indicative of fear, one is already familiar. The behaviour that reduces distance from persons or objects that are treated as though they provided protection is nothing other than attachment behaviour. Viewed in this perspective,
-89-
therefore, though not in others, attachment behaviour appears as one component among the heterogeneous forms of behaviour commonly grouped together as fear behaviour.
It is evident that, if confusion is to be avoided, distinctive names are required also for any other components of fear behaviour that can be clearly identified. For behaviour that tends to increase distance from persons and objects that are treated as though they were threatening, the terms 'withdrawal', 'escape', and 'avoidance' are all convenient. For another principal and well-organized component, namely behaviour that results in immobility, the usual term is 'freezing'. Since, in humans, freezing has been little studied, most of the coming discussion hinges on the relations between attachment behaviour and withdrawal behaviour.
There need, of course, be no surprise that attachment behaviour and withdrawal behaviour are so frequently found together. For, as already argued in the first volume, both serve the same function, namely protection; and, because of that, both show many of the same eliciting conditions. Furthermore, when they are active together, as so frequently they are, the two forms of behaviour are usually compatible: more often than not it is easy to combine in a single action withdrawing from one zone and approaching another. It is in fact for these very reasons that the two are so habitually paired together, without distinction or thought, under the general rubric of fear behaviour.
Nevertheless, although attachment behaviour and withdrawal behaviour have so much in common, there are strong reasons for keeping them distinct. One is that, although they share many eliciting conditions, they do not share all. For example, attachment behaviour may be activated by fatigue or illness as well as by a situation that arouses fear. Another is that when the two forms of behaviour are active together, though they are usually compatible, they may not be. Conflict can easily occur, for example, whenever a stimulus situation that elicits both escape and attachment behaviour in an individual happens to be situated between that individual and his attachment figure; a familiar instance is when a barking dog comes between a child and his mother.
In a conflict situation of that sort there are at least four ways in which the frightened individual may behave, depending on whether escape behaviour or attachment behaviour takes precedence or whether they are evenly balanced. Examples of
-90-
balance are when the frightened individual stays stationary, and also when he gets to his attachment figure by making a de? tour to avoid whatever is frightening him. Examples of one
78
or other form of behaviour taking precedence are when the frightened individual goes more or less directly to his attachment figure despite having to pass close to the frightening object in order to do so, and also when he runs away from the frightening object even though by doing so he increases the distance from his attachment figure.
Although there is a large literature on approach/avoidance conflict, it is doubtful whether any experiments have been undertaken to determine, in the case of this version of the conflict, which of these different solutions is favoured by creatures of different age and species, and in different conditions. Any assumption that escape behaviour commonly takes precedence over attachment would, however, certainly be wrong. Much everyday experience shows that, in young animals of many species, attachment behaviour frequently takes precedence over escape. An example is the behaviour of lambs on a hill road when a car approaches. Caught on the side of the road opposite to its mother and frightened by the approaching car, a lamb will as often as not rush across the road in front of the car. Small children are apt to do the same.
Studies of human behaviour during and after a disaster contain countless vivid accounts of how no member of a family is content, or indeed able to attend to anything else, until all members of the family are gathered together. The studies describe also the tremendous comfort that the presence of another familiar person can bring and how, during the weeks after a disaster, the rule is for people to remain in close contact with attachment figures. Again and again attachment behaviour takes precedence over withdrawal. The findings of some of these studies are referred to again at the end of Chapter 10.
A special but not unusual situation in which there is conflict between attachment behaviour and withdrawal is when the attachment figure is also the one who elicits fear, perhaps by threats or violence. In such conditions young creatures, whether human or non-human, are likely to cling to the threatening or hostile figure rather than run away from him or her (for references, see Volume I, Chapter 12). This propensity may be playing a part in so-called phobic patients, whose inability to leave home is found often to be a response to alarming threats made by their parents (see Chapters 18 and 19).
-91-
This analysis shows that attachment behaviour and withdrawal behaviour are distinct behavioural systems that (a) have the same function, (b) may be elicited by many of the same conditions, (c) are frequently compatible with each other, but (d) can easily be in conflict. In cases of conflict it is a matter for inquiry to discover which, if either, takes precedence.
Fear and Attack
Stimulus situations that are likely to arouse fear in humans can also, when circumstances are a little changed, evoke attack. The close link between the two very different forms of behaviour is considered in Chapter 8 in so far as it is seen and studied in animals and in Chapter 17 in so far as it occurs in humans.
Feeling afraid and its variants: feeling alarmed and feeling anxious
Whether compatible or in conflict with one another, attachment behaviour and escape behaviour are commonly elicited by many of the same stimulus situations and are, it is held,
79
always serving the same function, that of protection. It is not surprising, therefore, that in at least some circumstances the two forms of behaviour are accompanied by rather similar subjective experience. When confronted by a stimulus situation that makes us want to withdraw or escape from it, we are likely to describe ourselves as feeling afraid, or frightened, or alarmed, or perhaps anxious. Equally, whenever our attachment behaviour is aroused, perhaps by a similar sort of situation, but for some reason we are unable to find or reach our attachment figure, we are likely to describe how we feel in much the same words. For example, we might say, 'I was afraid you were gone', or 'I was frightened when I could not find you', or 'Your long absence made me anxious'.
This rather promiscuous use of language is both revealing and confusing. On the one hand, it strongly suggests that escape behaviour and attachment behaviour may share certain basic features in common. On the other, it becomes easy for the unwary to assume that, because in common speech words are used without discrimination, whatever is referred to can be treated as though it were undifferentiated. In addition, promiscuous usage makes it extremely difficult to tie any specialized meaning to any particular word.
It has already been emphasized how, despite Freud's increasing insistence on the key role of separation anxiety in
-92-
neurosis, there has been marked reluctance to adopt his ideas, partly because of the influence of his earlier theories and partly because of the difficulty that both he and others have had in understanding why separation should in and of itself engender fear or anxiety. This long- lasting difficulty is well illustrated in a passage in a recent book on anxiety by Rycroft ( 1968a), comment on which serves to further the argument.
After referring briefly to evidence of the kind set out fully in Chapters 3 and 4 of this volume, Rycroft proceeds:
Observations of this kind, made on both animals and human infants, have given rise to the idea that all anxiety -- or at least all neurotic anxiety -- is in the last resort separationanxiety, a response to separation from a protecting, parental object rather than a reaction to unidentified danger. There are, however, objections to this idea. In the first place it is surely illogical to regard the absence of a known, protective figure rather than the presence of an unknown, threatening situation as the cause of anxiety. To do so is like attributing . . . frostbite to inadequate clothing and not to exposure to extreme cold.
Reflection shows that there is in fact nothing illogical in making the attributions to which Rycroft objects. The causal conditions producing frostbite include both extreme cold and inadequate clothing. It is, therefore, just as reasonable to inculpate the one as the other. 1
For our purpose, however, another analogy in which two conditions are equally relevant for safety is more apt. The safety of an army in the field is dependent not only on its defending itself against direct attack but also on its maintaining open communications with its base. Any military commander who fails to give as much attention to his base and lines of
____________________
1 Rycroft puts forward two other arguments to support his case. One is that 'the young of
80
both animals and man do not invariably become anxious when left alone; they may remain quiet and contented unless some other disturbing element is present'. This argument has substance and is discussed in Chapter 12. The other is that 'the exposure of infants and young animals to simultaneous stress and isolation is an unnatural artefact'. This is certainly not so. There is ample evidence that exposure of infants and young animals to simultaneous stress and isolation occurs in the wild, even if infrequently (see, for example, van Lawick-Goodall's observations on young chimpanzees, described briefly in Chapter 4, p. 59 above).
-93-
communication as to his main front soon finds himself defeated. The thesis advanced here, then, is 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.
Though a military analogy is useful, it requires amplification. As a rule a commander-in-chief in charge of front-line forces is also in command of his base. Therefore any threat to his base or to his lines of communication is likely to come only from a single source, the enemy. Let us suppose, by contrast, that the general commanding the front-line forces is not in command of the base, and that another general of equal or superior status is in charge there. In such a situation the general commanding at the front could well have two sources of anxiety, one regarding possible enemy attack and the other regarding possible defection by his colleague at base. Only if there were complete confidence between the two generals could the arrangement be expected to work.
A situation of that kind, it is suggested, holds between an individual and his attachment figure. Each party is inherently autonomous. Given basic trust the arrangement can work well. But any possibility of defection by the attachment figure can give rise to acute anxiety in the attached. And should he be experiencing alarm from another source at the same time, it is evident that he is likely to feel the most intense fear.
In clinical work, it is held, we should be as much concerned with threats to rear as with threats to front. In Part III of this volume evidence is presented that suggests that the acute and chronic anxieties of patients stem as often from breakdown in relations with base as from all other hazards put together. It is, indeed, a special merit of some psychoanalytic traditions that, in their concern with object relations, they have focused attention especially on relations with base.
It is necessary to emphasize that at one important point the military analogy breaks down. Whereas generals are concerned to assess real dangers, animals and children, and in great degree human adults also, are attuned to respond mainly to rather simple stimulus situations that act as natural clues either to an increased risk of danger or to potential safety, clues that are only roughly correlated with actual danger or actual safety. This much neglected fact was touched on at the end of the previous chapter and is explored systematically in Chapters 8, 9, and 10.
-94-
81
Terminology
The fact that the same vocabulary is in everyday use to describe how we feel both when threatened with attack and when our base is threatened suggests a similarity of feeling in the two situations. Yet it seems probable that feeling experience in the two situations is not identical. For this reason there would be advantages in having distinctive words.
In discussion of the problem in earlier papers ( Bowlby 1960a; 1961a) and again very briefly in the first volume (end of Chapter 15) a usage is proposed not dissimilar to that adopted by Freud in his later work. In so far as we may try at times to withdraw or escape from a situation, the word 'alarmed' is in many ways a suitable one to describe how we feel. In so far as we may at times be seeking an attachment figure but be unable to find or reach him (or her), the word 'anxious' is in many ways a suitable one to describe how we feel. This usage can be supported by reference both to the etymological roots of the respective words and to psychoanalytic tradition. Supporting arguments are presented in Appendix III to this volume.
In the terminology adopted, therefore, fear behaviour and feeling afraid are used as general- purpose terms, terms that encompass all forms of behaviour and, for humans, all shades of feeling also. When greater discrimination is required, the terms used are freezing and withdrawal or escape behaviour, which go with feeling alarmed, and attachment behaviour, which, when not terminated, goes with feeling anxious. Not infrequently, of course, a person is trying simultaneously to escape from one situation and, without success, to gain proximity to another. In such a case he would be described in this terminology as feeling both alarmed and anxious.
-95-
Chapter 7 Situations that Arouse Fear in Humans
. . . certain ideas of supernatural agency, associated with real circumstances, produce a peculiar kind of horror. This horror is probably explicable as the result of a combination of simpler horrors. To bring the ghostly terror to its maximum, many usual elements of the dreadful must combine, such as loneliness, darkness, inexplicable sounds, especially of a dismal character, moving figures half-discerned . . . , and a vertiginous baffling of the expectation. This last element, which is intellectual, is very important.
WILLIAM JAMES ( 1890)
A difficult field of study
Evidence is given (Chapters 3 and 4) of the distress and anxiety that are aroused when young creatures, human and other, are removed from a figure to whom they are attached and placed with strangers. In such circumstances, we know, behaviour is directed as least as much to regaining the familiar figure as it is to escaping from the strange people and situation. In those chapters attention is concentrated on the effects on behaviour of a single variable, presence or absence of mother; and thereby light is thrown on one-half of our problem, a half hitherto
82
much neglected. It is time now to give attention to the other and more familiar half, the nature of some of the other variables that are likely to elicit one or another form of fear behaviour.
Not only are the forms of behaviour usually classified as fear behaviour heterogeneous, but so also, as we have seen, are the immediate situations and events that commonly elicit them. They include, besides being lost or alone, sudden noises and movements, strange objects and persons, animals, height, rapid approach, darkness, and anything we have learnt can cause pain. It is a motley list. And not only motley but the power of each situation or event to elicit fear is most uncertain. Where one person is afraid another is not. Where someone is unafraid today he is afraid tomorrow, or the other way round.
To all these immediate and concrete situations that are apt
-96-
to arouse fear must then be added all those potential situations that, on good or less good grounds, a person may foresee as disagreeable or dangerous, including all the so-called imaginary fears.
The resulting scene is undoubtedly confusing; and perhaps it is no wonder that, in the attempt to understand it, many theories have been advanced, some empirically based, some more speculative, some testable and others not. At one extreme is J. B. Watson's simplistic theory that stimulus situations of any kind that later elicit fear can be traced to primal fear of two basic stimulus situations, one a loud sound and the other loss of support; at another is the type of theory first put forward by Freud, and carried further by some of his followers, that regards the situations a man fears in the external world as being reflections mainly of the danger situations he encounters in his internal one.
But we need not be tossed hither and yon. When the empirical evidence is arrayed, derived from studies both of men and of animal species, not only do the characteristics of feararousing situations become clear but the contribution made to species survival by response to them is usually not difficult to see. A finding of central importance to the argument is that two stimulus situations that, when present singly, might arouse fear at only low intensity may, when present together, arouse it at high intensity. Another and related one is that the presence or absence of an attachment figure, or other companion, makes an immense difference to the intensity of fear aroused. Only if these two findings are borne constantly in mind can the conditions that elicit intense fear be understood.
In this chapter an account is given of the situations that commonly arouse fear in humans, and in the next a comparable account of the situations that commonly arouse fear in animals. In each chapter the earlier part is concerned with stimulus situations that seem to have an inherent potential for arousing fear and eliciting one or another form of fear behaviour, and the later part with the greatly intensified effects seen when an individual is confronted by a situation compounded of two or more of such fear-inducing situations, including being alone.
Considering the immense importance of fear in human life, and especially in psychiatric illness, it is surprising how few researchers have made systematic attempts to study the situations that commonly arouse fear in humans. Very recently, it is
-97-
83
true, a fresh start has been made with the empirical investigation of situations that arouse fear during the first year or so of life. During that phase experiment is not too difficult because both mobility and cognitive development are limited. Once a child is beyond that phase, however, conditions for studying fear grow harder. Very few first-hand studies have been reported and, instead, there has been a tendency to rely on reports given by mothers during interview. Though such reports are of some value, there are several reasons why their value is limited.
Inadequacy of Mothers' Reports
Mothers are not expert observers nor are they disinterested. As will appear, the study of situations that give rise to fear is technically very difficult. First, it must be agreed what forms of behaviour are, and what are not, to count as indicative of fear. Next, it becomes evident that whether fear behaviour is shown is enormously influenced both by particular environmental conditions and by the state of the child: unless these details are reported, interpretation of results is difficult or impossible.
Apart from the technical difficulties in reporting, no mother is disinterested, and some may be heavily biased. A mother may exaggerate or minimize the intensity of her child's fear responses, or overlook or invent situations that elicit fear in him. In such matters the possibility of wishful thinking or of attributing to her child fears that belong only to herself is obvious. Another difficulty is that inevitably a mother is often ignorant of what does and what does not make her child afraid.
Marked discrepancies in reports made independently by mothers and their children were found in a study by Lapouse & Monk ( 1959). A sample of 193 children aged between eight and twelve years were interviewed and asked about what situations made them afraid; the mothers were also interviewed, separately, and asked the same questions. Disagreement between informants varied from a mere 7 per cent in regard to certain situations to as high as 59 per cent in regard to others. When mother and child gave different answers the reason was very frequently that a child described himself as being afraid of a situation of which his mother said he was not afraid. Among situations notably under-reported by mothers were the following: fear of getting lost or kidnapped; fear of strangers; fear of calamities such as fire, wars, floods, and murders; fear of a member of the family falling sick, having an accident, or
-98-
dying; fear of falling sick oneself. 1 For each of these situations there were between 42 and 57 per cent of families in which mother reported her child as unafraid and the child reported otherwise. By contrast, in respect of these situations there were never more than 10 per cent of families in which the child claimed he was not afraid when his mother reported he was.
For all these reasons it is necessary to be very cautious in accepting the reports of mothers. In regard to the classes of situation likely to be feared their answers are of use. For calculating the proportion of a particular sample of children who are in fact prone to be afraid of a particular situation, their answers are insufficiently reliable. In what follows, therefore, we are guided mainly by results obtained from direct observation of children or from interviews with them.
84
Psychoanalysts and ethologists are agreed that a principal key to the understanding of any sort of behaviour is to study it developmentally. Nowhere is this perspective more necessary than in the study of fear behaviour in man. We start, therefore, with fear in infancy.
Fear-arousing situations: the first year
Initially during infancy the responses in which we are interested consist of little more than startle, crying, and diffuse movements. Whether it is useful to term them fear is almost a matter of taste. Because during the first three months there is so little discriminated perception or organized movement Bronson ( 1968) suggests they are better termed 'distress'. A little later, between the fourth and sixth months, a time when perceptual ability is developing, Bronson (in press) suggests it is useful to speak of an infant's being 'wary'.
During the second half of the first year, when perception becomes more discriminating and responses are better organized, the term fear is clearly appropriate. With greater or less efficiency an infant is effecting movement away from certain types of object or event, and towards others. By the end of the year, moreover, a baby can predict unpleasant happenings from the presence of simple clues that he has learnt. And during the second and, more especially, later years the capacity to
____________________
1 The incidence of fear of sickness etc. (either in a family member or in the self) reported in
this study is much higher than it is in others. The probable reason for this is that the special sample from which these findings come was drawn 'from the outpatient clinics of two hospitals and the offices of several pediatricians'.
-99-
foresee disagreeable situations and to take precautionary measures is greatly extended.
Early Situations and Responses
Bronson ( 1968) has reviewed studies of the types of stimulus situation that evoke distress responses during the early months of life.
Initially, discomfort, pain, and sudden sharp sounds upset a baby and may lead to crying, muscle tension, and diffuse movement.
By contrast, a baby is quieted by being rocked or patted and by engaging in non-nutritive sucking (see Volume I, Chapter 14). Although it has been thought that during the earliest months vision plays only a small part in arousing fear, a recent experiment ( Bower, Broughton & Moore 1970) shows that a baby of a few weeks flinches and cries whenever he sees an object approaching close to him. From about four months onwards, moreover, an infant is beginning to distinguish the strange from the familiar and to become wary of whatever is not familiar. Then, in a few children from about seven months, and in most by nine or ten, the sight of a stranger may be seen to arouse an unmistakable fear response. Some discussion of this response is already given in Volume I (Chapter 15). Since then further study of the genesis of the response has been undertaken by Bronson and by Scarr & Salapatek in the United States and by Schaffer in the United Kingdom. When account is taken of the different experimental situations and different methods of scoring responses, findings are highly compatible.
85
Fear of Strangers
Using videotape-recording and sensitive measures of response, Bronson (in press) has studied reactions to strange persons in thirty-two infants in their familiar home surroundings, during their development from the age of three to nine months. He reports that, usually starting at about four months, most infants will occasionally respond to a strange person with a cry, a whimper, or a frown, and that these wary responses start to appear at the age when undiscriminating smiling at strangers begins to wane. Throughout the fourth and fifth months, however, visual discrimination of strangers remains slow and uncertain. An infant of this age may spend long periods staring intently at a nearby stranger and may delay a long time before responding; and, on occasion, his response may change from a smile to a frown. Whether wariness is shown is determined by such
-100-
variables as the visual characteristics of the stranger, his proximity and manner of approach; but before six months, in contrast to later (see below, pp. 120 -2), it makes little difference whether or not an infant is held by, or can see, mother. At this age, moreover, the response in any one infant is far from stable.
Once infants are past six months, these responses usually become more differentiated and, for any one infant, more predictable. In the first place the response is more clearly aversive, so that the term fear becomes more applicable. In the second, perceptual identification of a stranger is posing fewer difficulties. Even so, as Schaffer ( 1971) points out, the earliest occasions on which an infant shows fear of strangers are when mother is herself present and the infant, by looking to and fro, can compare the two figures. Only later is an infant able readily to make the comparison, from memory, when his mother is absent.
As an infant reaches the end of his first year, his responses become still more predictable, and he may show unease either towards a particular person or towards members of a particular sex.
A point emphasized in the first volume is that, in any one infant, fear of strangers varies greatly according to conditions. How far distant from the infant a stranger is, whether he approaches, and whether he touches the infant are all extremely important; so also is the distance the child is from his mother. The significance of these variables for an understanding of fear is discussed later in this chapter and in those that follow.
Fear of Strange Objects
At about the same age as a child is becoming afraid of strangers he is apt also to become afraid of the sight of novel situations and strange objects. For example, Meili ( 1959) found, during a longitudinal study, that many infants become afraid of a jack-in-the-box at about ten months. Support for this finding comes from Scarr & Salapatek ( 1970) who undertook a crosssectional study of fear responses as they occur in children between the ages of about five and eighteen months. At every age between nine and fourteen months rather more than a third of the children were frightened both by the jack-in-the-box test and by the approach of a mechanical dog. Fewer of the younger and of the older children were afraid of these situations.
86
Schaffer has investigated the development of responses to unfamiliar objects. In a series of experiments, Schaffer & Parry
-101-
( 1969; 1970) have shown that at the age of six months, although infants are fully capable of perceiving differences between a familiar object and an unfamiliar one, they none the less approach both kinds of object without any discrimination. From about eight months of age onwards, however, infants begin to show sharp discrimination. Thenceforward, whereas a familiar object is approached with confidence, an unfamiliar one is treated with caution: in the experiments some children merely stared at the unfamiliar object, others appeared as if frozen, while yet others were distressed and withdrew. Even when, after they had become familiar with the object, they began to touch it, they still did so only briefly and tentatively.
A particularly interesting observation reported by Schaffer ( 1971) is the way in which a one- year-old infant habitually turns to his mother when uncertain, whereas an infant aged six months does not. Two groups of infants, one aged six months and the other twelve months, were presented with an array of stimulus objects. Behind each infant sat his mother, with instructions to say and do nothing unless her infant became upset. Whereas the younger infants seemed to be spellbound by the objects in front of them and unaware that mother was immediately behind, the older infants turned frequently from the objects to mother and back again, apparently well able to keep mother in mind despite her being perceptually absent. Thus by twelve months of age an infant is capable of organized fear behaviour characterized typically by movement away from objects of one class and towards objects of another. The development in an infant, during the second half of the first year, of the capacity to turn to mother when he is frightened and to find comfort in her presence is described in greater detail in the final section of this chapter.
Certain other conditions that regularly arouse fear behaviour in human infants during the second half of the first year are constellations of visual stimuli that act as naturally occurring clues to the imminence of two dangers common in the wild: the danger of falling, and the danger of being attacked or overwhelmed by an object rapidly approaching.
Fear of the Visual Cliff
Walk & Gibson ( 1961) have described the behaviour of thirtysix infants, aged from six to fourteen months, all of whom could crawl, when tested on an apparatus known as the 'visual
-102-
cliff'. This consists of a board laid across a sheet of heavy glass, with a patterned material that is directly beneath the glass on one side and, by dropping vertically, is several feet below it on the other. The infant is placed on the centre of the board while his mother stands at one or other side, calling her child to come to her across the glass-covered table which, according to the side at which she is standing, appears to be either solid table or deep chasm. Since each mother alternated her position it was easy to determine whether or not an infant was afraid to venture across the glass-covered 'chasm'.
Of the thirty-six infants tested only three, all boys, crossed the chasm to get to mother. All the others refused: some cried, others backed away to avoid the chasm, others peered through the
87
glass or patted it. Yet, when mother was on the 'solid' side, most of the infants crawled quickly towards her. In the majority, then, discrimination was very evident.
Scarr & Salapatek ( 1970) repeated the experiment with their sample and found that the older the child the more likely was he to refuse to cross the chasm. Whereas of the children aged between seven and eleven months nearly half were willing to cross to get to mother, all those aged thirteen months and over refused to do so.
Since Walk & Gibson tested the young of many animal species on the visual cliff, it was possible for them to draw broad conclusions. Certainly in other species, and probably in humans as well, it is clear that fear on perception of clues indicative of height develops very early, and even when an infant has had no experience of falling. The perceptual cue that appears to trigger off avoidance behaviour is 'motion perspective', namely the differential motion of foreground and background produced by the infant's own actions. Compared with lambs and kids, all of which show reliable discrimination and accurate avoiding movements from the beginning, human infants are both less reliable in discriminating and more clumsy in movement. Nevertheless, a strong bias to avoid the chasm was evident in all but a small minority.
Fear of an Approaching Object (Looming)
Another stimulus condition that seems to elicit a natural fear reaction in human infants and to do so very early in life is a visual stimulus that expands rapidly, which is habitually interpreted by adults as indicating something rapidly approaching.
Many years ago Valentine ( 1930) noted that approach elicits
-103-
fear in young children. A small girl of fourteen months, he reported, showed great fear of a teddy bear whenever it was moved towards her but would pick it up and kiss it whenever it was still.
In recent times Boweret al. ( 1970) have shown that a defensive response to an approaching object occurs in infants as young as two weeks (provided they are alert in an upright or near upright position). In a study of over forty infants they report that, each time a soft object (foam-rubber cube 20 cm per side) approaches to within about eight inches of a baby's face, without touching him, the infant pulls his head backwards, puts his hands between his face and the object, and cries loudly. The closer the object comes, the louder the cry. Further tests show that, when the stimulus consists merely of a shadow rapidly expanding on a screen, the response is similar though less intense. By contrast, when an object moves away there is no response. In the next chapter it is seen that young rhesus monkeys behave in very similar ways.
The fear-eliciting properties of an approaching or looming object have probably been underestimated in the past; and it seems likely that, in some of the experiments on an infant's response to strangers and to novel objects, approach of the stranger or object has played a larger part in determining a fear response than the experimenters realized.
88
A stimulus condition related to looming and approach is darkness. During the first year of life, fear of darkness, which is common in later years, is not very evident. Even so, by the age of ten months, infants are more likely to leave mother to enter and explore a brightly illuminated room than a dimly lit one ( Rheingold & Eckerman 1970).
Fear of an Anticipated Situation
Yet another situation that arouses fear, observable towards the end of the first year but not earlier, is when a baby uses current clues to anticipate something unpleasant. Levy ( 1951) describes the behaviour of babies of different ages when they catch sight of a doctor preparing to repeat an injection first given a few weeks earlier. Before eleven months of age only a very occasional infant was observed to react with fear. At ages eleven and twelve months, however, one-quarter of the sample did so. In such cases, it seems probable, learning from experience has occurred.
-104-
Thus by the end of the first year an infant is withdrawing in an organized way when he perceives any of a number of stimulus situations that can be regarded as naturally occurring clues to potentially dangerous situations. Further, he has learnt a good deal about his perceptual world. Towards the familiar and the strange, and towards what he has learnt is agreeable and what disagreeable, he is behaving with a rough-and-ready discrimination. He moves towards the one and away from the other.
Fear-arousing situations: the second and later years Sources of Data
It has already been noted how few researchers have made systematic attempts to study the situations that commonly arouse fear in humans. Most of the few data published during recent decades come from various longitudinal studies of children developing. Examples are a study by Macfarlane, Allen & Honzik ( 1954) of about one hundred children in California and another by Newson & Newson ( 1968) of 700 children and their parents in an English urban community. In none of these studies, however, was the nature of the situations that arouse fear at the centre of interest, nor was the information reported obtained either from direct observation or from interview of the children themselves. The latter limitation applies also to the findings of a cross-sectional study of nearly 500 children undertaken by Lapouse & Monk ( 1959) in New York State. 1 In all these projects information came only from mothers.
Because of the paucity of recently gathered data it is necessary to turn to the results of work done during the early years of child development research.
Some forty years ago an American psychologist, A. T. Jersild, began a series of studies in which he set out to describe the kinds of situation in which children exhibit fear, and how these change as a child grows older. 2 Different methods of obtaining information were employed in different studies. The four main methods were: day-to-day recording by parents; simple experiments; interviewing children about situations of which
____________________
1 A representative sample, distinct from the smaller sample referred to on p. 98 above.
2 Jersild's principal studies are published as monographs: Jersild Markey & Jersild ( 1933)
89
and Jersild & Holmes ( 1935a). Abstracts of these and other studies with full references are in the symposium on Child Behavior and Development edited by Barker, Kounin & Wright ( 1943), and also in Jersild ( 1947)
-105-
they were currently afraid; giving questionnaires to adults about what they recalled of the situations that made them afraid as children. The subjects of each study were different and were drawn from different age-ranges. Despite a number of shortcomings, these studies are still much the most extensive yet attempted and therefore remain principal sources. Not only do the findings conform with ordinary experience, but at a number of points they are supported and amplified both by earlier (e. g. Hagman 1932) and by more recent work.
Findings from Parents' Records and Naturalistic Observations
The aim of one of Jersild's studies was to obtain a detailed record of the occasions and situations in which ordinary children exhibit fear during the course of their everyday lives. To this end the parents of over one hundred young children were enlisted, all of whom were prepared to keep detailed records of every occasion on which their child showed fear during a period lasting twenty-one days. Mimeographed forms and instructions were issued. On each occasion that a child showed fear parents were asked to record: (a) the actual behaviour exhibited (e. g. startle, withdrawing, turning to an adult, cries or other vocalizations, and words spoken); (b) the situation in which the behaviour occurred, in terms not only of its apparent cause (specific stimulus) but of the setting (place, time, what child was doing, persons present); and (c) the child's current condition (well or ill, fresh or tired).
In all, 136 records were obtained of children between twelve and fifty-nine months of age (the records of a few younger and older children were too few to give useful results). The sample was biased towards the higher end of the socio-economic scale. A majority lived in a large city, but there were also children living in suburbs, small towns, and rural areas. Distribution by age was as follows: second year, twenty-three children; third year, forty-five; fourth year, forty-six; and fifth year, twenty-two children.
The authors were struck by how few were the occurrences of fear recorded by the parents of children in these age-groups, a finding confirmed when observations were made of some of the same children in a nursery school. The occurrences recorded during the three weeks for children in the two younger agegroups averaged only six per child, or two a week. For children in the two older age-groups the average was three and a half per child, or barely more than one a week. For about one in
-106-
ten of the children at each age-level no occurrence of fear was recorded throughout the three- week period. Although these figures suggest that some occurrences of slight or transient fear may have gone unrecorded, there is independent evidence that at least some children during their second year show fear very infrequently. For example, Valentine ( 1930), who kept daily notes on his children, was impressed by the infrequency with which he saw fear responses in them and describes how surprised he was when a child who had fallen and hurt himself would none the less immediately go climbing again. Anderson ( 1972a), who observed fifty-two children between the ages of twelve months and three years two months in a London park,
90
also remarks how infrequently the children exhibited fear. Occurrences, he reports, were 'uncommon and short-lived'. It should, however, be noted that in both these studies and also in Jersild's the children observed were not alone. The difference made by presence or absence of a trusted adult cannot be exaggerated (see the final section of this chapter). When we turn to examine the kinds of situation that are reported to arouse fear we find remarkably little change between the second and the fifth years. From the mother's day-to-day records reported by Jersild six situations stand out as being likely to arouse fear at least occasionally in a fairly large proportion of the children at each age-level:
--noise, and events associated with noise
--heights
--strange people or familiar people in strange guise --strange objects and surroundings
--animals
--pain or persons associated with pain.
In each of these six situations some 40 per cent of all the children were recorded as having shown fear behaviour on some occasion during the three-week spell. In so far as there was any reduction with age in the proportion of children showing fear the reduction occurred after the third birthday. 1
____________________
1 When analysed by age the picture from these six situations is as follows. Of the one-year-
olds, 60 per cent were recorded as having shown fear of noise, 52 per cent fear of pain or potential pain, and from 35 to 40 per cent as having shown fear in each of the other four situations, of which one was the presence of animals. Of the four-year-olds, only 23 per cent were recorded as having shown fear of noise and events
-107-
Among the many other situations recorded as having aroused fear, but in a smaller percentage of these children, were sudden unexpected movements, especially when an object was both approaching and noisy, and also bright lights, flashes, etc. Together, these kinds of situation elicited fear in nearly 30 per cent of the one- and two-year-olds; but of the older children no more than about 10 per cent had shown fear in those situations. The dark, especially being alone in the dark, was recorded as having aroused fear in about 10 per cent of the children during the three-week period of observation: here there was no change with age. Fear of being left alone or abandoned was recorded for about 10 per cent of the children at each of the age- levels. Only after the second birthday were any of the children recorded as having shown fear of imaginary creatures, the proportion being about 6 per cent. The origin and nature of such fears are discussed in Chapters 10 and 11.
The forms of behaviour recorded as shown by these children when afraid differed very little between the younger and the older ones. The behaviour most frequently recorded was crying in its various forms, from whimpering to screaming and including explicit cries for help. In each age-group no less than onethird of the episodes of fear recorded by mothers were signalled by crying of one sort or another. Another form of behaviour also frequently recorded was either turning towards an adult or running to him or her, with or without clinging: in each age-group about one-sixth of the fear-producing situations were said to elicit behaviour of that sort. Avoiding action or running away was recorded in about one-fifth of the episodes. In the remainder, fear was inferred because a child trembled or jumped, showed a frightened
91
expression, hid his head, or stayed unwontedly still. Occasionally a child was aggressive, or protective of another.
In this catalogue of forms of behaviour held by mothers to be indicative of fear it should be noted that two of the commonest are crying for or turning towards a protective figure. This finding resembles that of Anderson ( 1972a) when he interviewed the mothers of eighteen two-year-olds in London. The forms of fear behaviour most commonly described to him were screaming, crying, turning to mother, clinging to or
____________________
associated with noise, but no fewer than 40 per cent had shown fear of animals, the same percentage as for the one-year-olds. In each of the other situations, however, including the possibility of pain, only about 15 per cent of the four-year-olds had shown fear.
-108-
following mother, and staying close to her. Withdrawing from the frightening object was referred to less frequently.
When, however, Anderson was himself observing the fear behaviour of another but very similar sample of toddlers while they were with mother in a London park (see p. 107 above), rather different forms of fear behaviour were seen. In the dozen episodes observed, the fear- arousing object was an approaching animal (eight cases), an approaching child (three cases), and noise (one case). In such conditions a toddler would suddenly cease his activity, back away from whatever it was that was frightening him while continuing to fixate it, and simultaneously edge towards his mother. Crying was not observed. Once the object had moved away, the child would advance again but continue to fixate it.
The differences in forms of behaviour exhibited turn presumably on the intensity of fear. At high intensity crying and clinging are common; at low intensity withdrawing from the object and backing towards mother.
A limitation of records kept by parents and also of naturalistic observations such as those of Anderson is that, when a child is not described as having shown fear of a particular class of situation during the period his behaviour was recorded, it remains uncertain whether he is never afraid in such situations or whether it just happened that during that span of time he was never confronted by one when his mother or the observer was about. The experiments devised by Jersild & Holmes help to clarify this matter, though they too have obvious limitations.
Experimental Findings
It is evident that ethical considerations limit severely the kinds of experiment it is legitimate to do in order to explore the situations that arouse fear in humans, especially in young children.
In their experimental work with children between their second and sixth birthdays, therefore, Jersild & Holmes took many precautions. In the first place, each child was throughout with an adult who was experienced with children, and who had made a friendly contact with him before the experiments began. In the second, each situation with which a child was confronted
92
was one that many children hardly find frightening. In the third, a child was introduced to the situation by easy stages. And finally, if a child refused to take part, the experiment was ended.
-109-
There were eight potentially fear-provoking situations. Four were presented on the first day, extending over about fifteen minutes; two on the day following; and two about four weeks later. For a few minutes between each fear-inducing situation a child was allowed to play with toys. Since it appears that the situations were presented to every child in the same order, there is a serious possibility that responses to later situations may have been influenced by the experience of earlier ones; though in which direction such influence may have been exerted it is difficult to know. On the one hand is the possibility that, through habituation, responses to later situations may have been less than they would otherwise have been. On the other is the possibility that, as the series progressed, a child may have become increasingly sensitized and therefore have shown more fear in some of the later situations than he would otherwise have done. The finding that a larger proportion of children showed fear in the later experimental situations than in the earlier ones is in keeping with the latter possibility. The eight situations were chosen because earlier studies had shown that they were likely to elicit at least a little fear in a substantial proportion of young children. Details are as follows:
1. Being left alone: When the child is seated at a table playing with a toy, the experimenter names a pretext for leaving the room (which until the time of the experiment was unfamiliar to the child). The experimenter remains outside the room for two minutes. The child's behaviour is recorded by concealed observers.
2. Sudden displacement or loss of support: A bridge-like piece of apparatus, consisting of two boards laid end to end at a height of about two inches above the floor, was used. The first board is securely supported, but when the child steps onto the second board, which is supported only at the middle, it gives way and descends to the floor.
3. Dark passage: While playing ball with the child, the experimenter seemingly inadvertently throws the ball into a dark passage, 18 ft long, leading from one corner of the room. The child is asked to retrieve the ball.
4. Strange person: While the child is temporarily withdrawn from the room an assistant, dressed in a long grey coat, a large black hat, and a veil that obscures her features, seats herself in one of the two chairs near the entrance. The child
-110-
returns and his reactions are observed when he notices the stranger and when he is asked to obtain toys placed near the stranger's chair.
5. High board: A board 12 in. wide, about 8 ft long, and 2 in. thick, held firmly in place at the ends by two stationary holders, is arranged at various heights from the ground and the child is asked to walk from one end of the board to the other to obtain a box of brightly coloured toys. The board is first placed at a distance of four feet from the floor; it is subsequently lowered if the child refuses to walk across at this height and is raised if he performs at the four-foot level.
6. Loud sound: An iron pipe 2 ft long and 2 1/4 in. diameter, suspended from the ceiling in a corner of the room behind a screen, is struck a sharp blow with a hammer while the child and the experimenter are seated at a table containing toys. The child's response to the unexpected noise from an unseen source is first observed and then, pointing to the screen, the experimenter asks the child to 'Go and see what made that noise'.
7. Snake: A snake, harmless and about 2 ft long, was placed in a box sufficiently deep to 93
ensure that it could not immediately climb out when the top was removed. In the box was placed a small coloured toy. The child's attention is directed to the box, the lid is uncovered, and the child is allowed to look in; if he raises any questions the experimenter simply says 'It is a snake', and then points to the toy and asks the child to reach in and get the toy.
8. Large dog: While the child is seated at a table with toys a large collie dog is brought into the room on a leash by a familiar person. The dog is led to a certain point in the room and, after preliminary comment by the experimenter, the child is asked to go and pat the dog.
The subjects were 105 children, half from a private nursery school for better-off families and half from a public nursery school for poorer families. Fifty-seven were boys and fortyeight girls. Children were tested only when they were fit, willing, and in a good humour; and the experiments were never combined with any other sort of examination. On each occasion when tests were carried out, two-year-olds and three-yearolds were well represented (never fewer than twenty-one children tested and usually between thirty and forty-five); but
-111-
four-year-olds and five-year-olds were rather few (numbers tested varied between seven and fourteen children). Each experiment, except the first, was presented to a child in four stages: first he was given directions what to do; then, if he was hesitant, he was given reassurance and encouragement; then, if he still refrained, the experimenter offered to accompany him in the task; finally, if a child was still reluctant to take part, the experiment was abandoned.
-87-
and how we predict he is likely to behave (see Volume I, Chapter 7). Since there is reason to think that fear behaviour has hitherto been given far too little attention, that is where we start. Fear Behaviour
Let us examine the various forms of behaviour that are commonly held to be indicative of fear. They include, of course, the initial forms of behaviour, such as posture, expression, and
76
incipient action, that lead us to infer that a person or animal is feeling afraid, and also the less subtle and more active forms of behaviour that often, but not always, follow. Both in ordinary life and in systematic field observation there is a large array of distinct forms of behaviour that it is common practice to group together as being indicative of fear. They include wary watching combined with inhibition of action, a frightened facial expression accompanied perhaps by trembling or crying, cowering, hiding, running away, and also seeking contact with someone and perhaps clinging to him or her. When we ask why all these diverse forms of behaviour should be grouped together, we find the following four reasons:
a. many of these forms of behaviour, though not all, tend to occur either simultaneously or sequentially;
b. events that elicit one of these forms tend to elicit others also (though not necessarily all the others);
c. most of them seem plainly to serve a single biological function, namely protection;
d. when asked how they are feeling, persons behaving in these ways commonly describe
themselves as feeling afraid or anxious or alarmed.
Although these are good reasons for grouping such diverse forms of behaviour together, there are nevertheless risks in doing so. In particular, the conditions that elicit one form of fear behaviour may differ in certain respects from those that elicit another form; and the autonomic responses that accompany one form may well differ from those that accompany another. In animals the distinctiveness of the forms is attested by experiment. Hinde ( 1970) discusses work by Hogan that suggests that, at least in young animals, not only may freezing and withdrawal be separate systems of behaviour, elicited by different types of external stimulation, but they may even be mutually inhibiting. Further evidence on this issue is given in Chapter 8.
-88-
A point to be especially noted, and a corner-stone of the present argument, is that, in ordinary usage, grouped under this single heading of behaviour indicative of fear are to be found forms of behaviour that have at least three distinct kinds of predictable outcome: (a) immobility, (b) increased distance from one type of object, and (c) increased proximity to another type of object. The contrast between the last two outcomes is especially important. For, on the one hand, is behaviour that increases distance from persons and objects that are treated as though they were threatening; on the other is behaviour that reduces distance from persons and objects that are treated as though they provided protection. Naturally, both these types of behaviour do not always occur. Yet they occur together with sufficient frequency to enable us to take the combination for granted. When we flush a rabbit we expect it not only to run from us but to run to cover. When a child is afraid of a barking dog we expect him not only to withdraw from the dog but to retreat towards a parent figure.
Now the usual practice of including under a single heading, that of behaviour indicative of fear, forms of behaviour that have such different predictable outcomes is of great significance. Yet it can very easily make for confusion. In particular it has often tempted psychologists, for example McDougall ( 1923), and also others, to postulate a single all-embracing 'instinct of fear'. An alternative theory and one that keeps far closer to observed data is that we are dealing, not with some single comprehensive form of behaviour, but with a heterogeneous collection of interrelated forms, each elicited by a slightly different set of causal conditions and each having a distinctive outcome. In the sense defined in Part II of the first volume, each such form can be regarded as an example of instinctive behaviour.
77
In order to sort out these different forms of behaviour a first step is to examine how attachment behaviour and fear behaviour are related to one another.
Withdrawal behaviour and attachment behaviour
It may already have been noticed that, of the three forms of behaviour with such different predictable outcomes that are habitually treated as indicative of fear, one is already familiar. The behaviour that reduces distance from persons or objects that are treated as though they provided protection is nothing other than attachment behaviour. Viewed in this perspective,
-89-
therefore, though not in others, attachment behaviour appears as one component among the heterogeneous forms of behaviour commonly grouped together as fear behaviour.
It is evident that, if confusion is to be avoided, distinctive names are required also for any other components of fear behaviour that can be clearly identified. For behaviour that tends to increase distance from persons and objects that are treated as though they were threatening, the terms 'withdrawal', 'escape', and 'avoidance' are all convenient. For another principal and well-organized component, namely behaviour that results in immobility, the usual term is 'freezing'. Since, in humans, freezing has been little studied, most of the coming discussion hinges on the relations between attachment behaviour and withdrawal behaviour.
There need, of course, be no surprise that attachment behaviour and withdrawal behaviour are so frequently found together. For, as already argued in the first volume, both serve the same function, namely protection; and, because of that, both show many of the same eliciting conditions. Furthermore, when they are active together, as so frequently they are, the two forms of behaviour are usually compatible: more often than not it is easy to combine in a single action withdrawing from one zone and approaching another. It is in fact for these very reasons that the two are so habitually paired together, without distinction or thought, under the general rubric of fear behaviour.
Nevertheless, although attachment behaviour and withdrawal behaviour have so much in common, there are strong reasons for keeping them distinct. One is that, although they share many eliciting conditions, they do not share all. For example, attachment behaviour may be activated by fatigue or illness as well as by a situation that arouses fear. Another is that when the two forms of behaviour are active together, though they are usually compatible, they may not be. Conflict can easily occur, for example, whenever a stimulus situation that elicits both escape and attachment behaviour in an individual happens to be situated between that individual and his attachment figure; a familiar instance is when a barking dog comes between a child and his mother.
In a conflict situation of that sort there are at least four ways in which the frightened individual may behave, depending on whether escape behaviour or attachment behaviour takes precedence or whether they are evenly balanced. Examples of
-90-
balance are when the frightened individual stays stationary, and also when he gets to his attachment figure by making a de? tour to avoid whatever is frightening him. Examples of one
78
or other form of behaviour taking precedence are when the frightened individual goes more or less directly to his attachment figure despite having to pass close to the frightening object in order to do so, and also when he runs away from the frightening object even though by doing so he increases the distance from his attachment figure.
Although there is a large literature on approach/avoidance conflict, it is doubtful whether any experiments have been undertaken to determine, in the case of this version of the conflict, which of these different solutions is favoured by creatures of different age and species, and in different conditions. Any assumption that escape behaviour commonly takes precedence over attachment would, however, certainly be wrong. Much everyday experience shows that, in young animals of many species, attachment behaviour frequently takes precedence over escape. An example is the behaviour of lambs on a hill road when a car approaches. Caught on the side of the road opposite to its mother and frightened by the approaching car, a lamb will as often as not rush across the road in front of the car. Small children are apt to do the same.
Studies of human behaviour during and after a disaster contain countless vivid accounts of how no member of a family is content, or indeed able to attend to anything else, until all members of the family are gathered together. The studies describe also the tremendous comfort that the presence of another familiar person can bring and how, during the weeks after a disaster, the rule is for people to remain in close contact with attachment figures. Again and again attachment behaviour takes precedence over withdrawal. The findings of some of these studies are referred to again at the end of Chapter 10.
A special but not unusual situation in which there is conflict between attachment behaviour and withdrawal is when the attachment figure is also the one who elicits fear, perhaps by threats or violence. In such conditions young creatures, whether human or non-human, are likely to cling to the threatening or hostile figure rather than run away from him or her (for references, see Volume I, Chapter 12). This propensity may be playing a part in so-called phobic patients, whose inability to leave home is found often to be a response to alarming threats made by their parents (see Chapters 18 and 19).
-91-
This analysis shows that attachment behaviour and withdrawal behaviour are distinct behavioural systems that (a) have the same function, (b) may be elicited by many of the same conditions, (c) are frequently compatible with each other, but (d) can easily be in conflict. In cases of conflict it is a matter for inquiry to discover which, if either, takes precedence.
Fear and Attack
Stimulus situations that are likely to arouse fear in humans can also, when circumstances are a little changed, evoke attack. The close link between the two very different forms of behaviour is considered in Chapter 8 in so far as it is seen and studied in animals and in Chapter 17 in so far as it occurs in humans.
Feeling afraid and its variants: feeling alarmed and feeling anxious
Whether compatible or in conflict with one another, attachment behaviour and escape behaviour are commonly elicited by many of the same stimulus situations and are, it is held,
79
always serving the same function, that of protection. It is not surprising, therefore, that in at least some circumstances the two forms of behaviour are accompanied by rather similar subjective experience. When confronted by a stimulus situation that makes us want to withdraw or escape from it, we are likely to describe ourselves as feeling afraid, or frightened, or alarmed, or perhaps anxious. Equally, whenever our attachment behaviour is aroused, perhaps by a similar sort of situation, but for some reason we are unable to find or reach our attachment figure, we are likely to describe how we feel in much the same words. For example, we might say, 'I was afraid you were gone', or 'I was frightened when I could not find you', or 'Your long absence made me anxious'.
This rather promiscuous use of language is both revealing and confusing. On the one hand, it strongly suggests that escape behaviour and attachment behaviour may share certain basic features in common. On the other, it becomes easy for the unwary to assume that, because in common speech words are used without discrimination, whatever is referred to can be treated as though it were undifferentiated. In addition, promiscuous usage makes it extremely difficult to tie any specialized meaning to any particular word.
It has already been emphasized how, despite Freud's increasing insistence on the key role of separation anxiety in
-92-
neurosis, there has been marked reluctance to adopt his ideas, partly because of the influence of his earlier theories and partly because of the difficulty that both he and others have had in understanding why separation should in and of itself engender fear or anxiety. This long- lasting difficulty is well illustrated in a passage in a recent book on anxiety by Rycroft ( 1968a), comment on which serves to further the argument.
After referring briefly to evidence of the kind set out fully in Chapters 3 and 4 of this volume, Rycroft proceeds:
Observations of this kind, made on both animals and human infants, have given rise to the idea that all anxiety -- or at least all neurotic anxiety -- is in the last resort separationanxiety, a response to separation from a protecting, parental object rather than a reaction to unidentified danger. There are, however, objections to this idea. In the first place it is surely illogical to regard the absence of a known, protective figure rather than the presence of an unknown, threatening situation as the cause of anxiety. To do so is like attributing . . . frostbite to inadequate clothing and not to exposure to extreme cold.
Reflection shows that there is in fact nothing illogical in making the attributions to which Rycroft objects. The causal conditions producing frostbite include both extreme cold and inadequate clothing. It is, therefore, just as reasonable to inculpate the one as the other. 1
For our purpose, however, another analogy in which two conditions are equally relevant for safety is more apt. The safety of an army in the field is dependent not only on its defending itself against direct attack but also on its maintaining open communications with its base. Any military commander who fails to give as much attention to his base and lines of
____________________
1 Rycroft puts forward two other arguments to support his case. One is that 'the young of
80
both animals and man do not invariably become anxious when left alone; they may remain quiet and contented unless some other disturbing element is present'. This argument has substance and is discussed in Chapter 12. The other is that 'the exposure of infants and young animals to simultaneous stress and isolation is an unnatural artefact'. This is certainly not so. There is ample evidence that exposure of infants and young animals to simultaneous stress and isolation occurs in the wild, even if infrequently (see, for example, van Lawick-Goodall's observations on young chimpanzees, described briefly in Chapter 4, p. 59 above).
-93-
communication as to his main front soon finds himself defeated. The thesis advanced here, then, is 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.
Though a military analogy is useful, it requires amplification. As a rule a commander-in-chief in charge of front-line forces is also in command of his base. Therefore any threat to his base or to his lines of communication is likely to come only from a single source, the enemy. Let us suppose, by contrast, that the general commanding the front-line forces is not in command of the base, and that another general of equal or superior status is in charge there. In such a situation the general commanding at the front could well have two sources of anxiety, one regarding possible enemy attack and the other regarding possible defection by his colleague at base. Only if there were complete confidence between the two generals could the arrangement be expected to work.
A situation of that kind, it is suggested, holds between an individual and his attachment figure. Each party is inherently autonomous. Given basic trust the arrangement can work well. But any possibility of defection by the attachment figure can give rise to acute anxiety in the attached. And should he be experiencing alarm from another source at the same time, it is evident that he is likely to feel the most intense fear.
In clinical work, it is held, we should be as much concerned with threats to rear as with threats to front. In Part III of this volume evidence is presented that suggests that the acute and chronic anxieties of patients stem as often from breakdown in relations with base as from all other hazards put together. It is, indeed, a special merit of some psychoanalytic traditions that, in their concern with object relations, they have focused attention especially on relations with base.
It is necessary to emphasize that at one important point the military analogy breaks down. Whereas generals are concerned to assess real dangers, animals and children, and in great degree human adults also, are attuned to respond mainly to rather simple stimulus situations that act as natural clues either to an increased risk of danger or to potential safety, clues that are only roughly correlated with actual danger or actual safety. This much neglected fact was touched on at the end of the previous chapter and is explored systematically in Chapters 8, 9, and 10.
-94-
81
Terminology
The fact that the same vocabulary is in everyday use to describe how we feel both when threatened with attack and when our base is threatened suggests a similarity of feeling in the two situations. Yet it seems probable that feeling experience in the two situations is not identical. For this reason there would be advantages in having distinctive words.
In discussion of the problem in earlier papers ( Bowlby 1960a; 1961a) and again very briefly in the first volume (end of Chapter 15) a usage is proposed not dissimilar to that adopted by Freud in his later work. In so far as we may try at times to withdraw or escape from a situation, the word 'alarmed' is in many ways a suitable one to describe how we feel. In so far as we may at times be seeking an attachment figure but be unable to find or reach him (or her), the word 'anxious' is in many ways a suitable one to describe how we feel. This usage can be supported by reference both to the etymological roots of the respective words and to psychoanalytic tradition. Supporting arguments are presented in Appendix III to this volume.
In the terminology adopted, therefore, fear behaviour and feeling afraid are used as general- purpose terms, terms that encompass all forms of behaviour and, for humans, all shades of feeling also. When greater discrimination is required, the terms used are freezing and withdrawal or escape behaviour, which go with feeling alarmed, and attachment behaviour, which, when not terminated, goes with feeling anxious. Not infrequently, of course, a person is trying simultaneously to escape from one situation and, without success, to gain proximity to another. In such a case he would be described in this terminology as feeling both alarmed and anxious.
-95-
Chapter 7 Situations that Arouse Fear in Humans
. . . certain ideas of supernatural agency, associated with real circumstances, produce a peculiar kind of horror. This horror is probably explicable as the result of a combination of simpler horrors. To bring the ghostly terror to its maximum, many usual elements of the dreadful must combine, such as loneliness, darkness, inexplicable sounds, especially of a dismal character, moving figures half-discerned . . . , and a vertiginous baffling of the expectation. This last element, which is intellectual, is very important.
WILLIAM JAMES ( 1890)
A difficult field of study
Evidence is given (Chapters 3 and 4) of the distress and anxiety that are aroused when young creatures, human and other, are removed from a figure to whom they are attached and placed with strangers. In such circumstances, we know, behaviour is directed as least as much to regaining the familiar figure as it is to escaping from the strange people and situation. In those chapters attention is concentrated on the effects on behaviour of a single variable, presence or absence of mother; and thereby light is thrown on one-half of our problem, a half hitherto
82
much neglected. It is time now to give attention to the other and more familiar half, the nature of some of the other variables that are likely to elicit one or another form of fear behaviour.
Not only are the forms of behaviour usually classified as fear behaviour heterogeneous, but so also, as we have seen, are the immediate situations and events that commonly elicit them. They include, besides being lost or alone, sudden noises and movements, strange objects and persons, animals, height, rapid approach, darkness, and anything we have learnt can cause pain. It is a motley list. And not only motley but the power of each situation or event to elicit fear is most uncertain. Where one person is afraid another is not. Where someone is unafraid today he is afraid tomorrow, or the other way round.
To all these immediate and concrete situations that are apt
-96-
to arouse fear must then be added all those potential situations that, on good or less good grounds, a person may foresee as disagreeable or dangerous, including all the so-called imaginary fears.
The resulting scene is undoubtedly confusing; and perhaps it is no wonder that, in the attempt to understand it, many theories have been advanced, some empirically based, some more speculative, some testable and others not. At one extreme is J. B. Watson's simplistic theory that stimulus situations of any kind that later elicit fear can be traced to primal fear of two basic stimulus situations, one a loud sound and the other loss of support; at another is the type of theory first put forward by Freud, and carried further by some of his followers, that regards the situations a man fears in the external world as being reflections mainly of the danger situations he encounters in his internal one.
But we need not be tossed hither and yon. When the empirical evidence is arrayed, derived from studies both of men and of animal species, not only do the characteristics of feararousing situations become clear but the contribution made to species survival by response to them is usually not difficult to see. A finding of central importance to the argument is that two stimulus situations that, when present singly, might arouse fear at only low intensity may, when present together, arouse it at high intensity. Another and related one is that the presence or absence of an attachment figure, or other companion, makes an immense difference to the intensity of fear aroused. Only if these two findings are borne constantly in mind can the conditions that elicit intense fear be understood.
In this chapter an account is given of the situations that commonly arouse fear in humans, and in the next a comparable account of the situations that commonly arouse fear in animals. In each chapter the earlier part is concerned with stimulus situations that seem to have an inherent potential for arousing fear and eliciting one or another form of fear behaviour, and the later part with the greatly intensified effects seen when an individual is confronted by a situation compounded of two or more of such fear-inducing situations, including being alone.
Considering the immense importance of fear in human life, and especially in psychiatric illness, it is surprising how few researchers have made systematic attempts to study the situations that commonly arouse fear in humans. Very recently, it is
-97-
83
true, a fresh start has been made with the empirical investigation of situations that arouse fear during the first year or so of life. During that phase experiment is not too difficult because both mobility and cognitive development are limited. Once a child is beyond that phase, however, conditions for studying fear grow harder. Very few first-hand studies have been reported and, instead, there has been a tendency to rely on reports given by mothers during interview. Though such reports are of some value, there are several reasons why their value is limited.
Inadequacy of Mothers' Reports
Mothers are not expert observers nor are they disinterested. As will appear, the study of situations that give rise to fear is technically very difficult. First, it must be agreed what forms of behaviour are, and what are not, to count as indicative of fear. Next, it becomes evident that whether fear behaviour is shown is enormously influenced both by particular environmental conditions and by the state of the child: unless these details are reported, interpretation of results is difficult or impossible.
Apart from the technical difficulties in reporting, no mother is disinterested, and some may be heavily biased. A mother may exaggerate or minimize the intensity of her child's fear responses, or overlook or invent situations that elicit fear in him. In such matters the possibility of wishful thinking or of attributing to her child fears that belong only to herself is obvious. Another difficulty is that inevitably a mother is often ignorant of what does and what does not make her child afraid.
Marked discrepancies in reports made independently by mothers and their children were found in a study by Lapouse & Monk ( 1959). A sample of 193 children aged between eight and twelve years were interviewed and asked about what situations made them afraid; the mothers were also interviewed, separately, and asked the same questions. Disagreement between informants varied from a mere 7 per cent in regard to certain situations to as high as 59 per cent in regard to others. When mother and child gave different answers the reason was very frequently that a child described himself as being afraid of a situation of which his mother said he was not afraid. Among situations notably under-reported by mothers were the following: fear of getting lost or kidnapped; fear of strangers; fear of calamities such as fire, wars, floods, and murders; fear of a member of the family falling sick, having an accident, or
-98-
dying; fear of falling sick oneself. 1 For each of these situations there were between 42 and 57 per cent of families in which mother reported her child as unafraid and the child reported otherwise. By contrast, in respect of these situations there were never more than 10 per cent of families in which the child claimed he was not afraid when his mother reported he was.
For all these reasons it is necessary to be very cautious in accepting the reports of mothers. In regard to the classes of situation likely to be feared their answers are of use. For calculating the proportion of a particular sample of children who are in fact prone to be afraid of a particular situation, their answers are insufficiently reliable. In what follows, therefore, we are guided mainly by results obtained from direct observation of children or from interviews with them.
84
Psychoanalysts and ethologists are agreed that a principal key to the understanding of any sort of behaviour is to study it developmentally. Nowhere is this perspective more necessary than in the study of fear behaviour in man. We start, therefore, with fear in infancy.
Fear-arousing situations: the first year
Initially during infancy the responses in which we are interested consist of little more than startle, crying, and diffuse movements. Whether it is useful to term them fear is almost a matter of taste. Because during the first three months there is so little discriminated perception or organized movement Bronson ( 1968) suggests they are better termed 'distress'. A little later, between the fourth and sixth months, a time when perceptual ability is developing, Bronson (in press) suggests it is useful to speak of an infant's being 'wary'.
During the second half of the first year, when perception becomes more discriminating and responses are better organized, the term fear is clearly appropriate. With greater or less efficiency an infant is effecting movement away from certain types of object or event, and towards others. By the end of the year, moreover, a baby can predict unpleasant happenings from the presence of simple clues that he has learnt. And during the second and, more especially, later years the capacity to
____________________
1 The incidence of fear of sickness etc. (either in a family member or in the self) reported in
this study is much higher than it is in others. The probable reason for this is that the special sample from which these findings come was drawn 'from the outpatient clinics of two hospitals and the offices of several pediatricians'.
-99-
foresee disagreeable situations and to take precautionary measures is greatly extended.
Early Situations and Responses
Bronson ( 1968) has reviewed studies of the types of stimulus situation that evoke distress responses during the early months of life.
Initially, discomfort, pain, and sudden sharp sounds upset a baby and may lead to crying, muscle tension, and diffuse movement.
By contrast, a baby is quieted by being rocked or patted and by engaging in non-nutritive sucking (see Volume I, Chapter 14). Although it has been thought that during the earliest months vision plays only a small part in arousing fear, a recent experiment ( Bower, Broughton & Moore 1970) shows that a baby of a few weeks flinches and cries whenever he sees an object approaching close to him. From about four months onwards, moreover, an infant is beginning to distinguish the strange from the familiar and to become wary of whatever is not familiar. Then, in a few children from about seven months, and in most by nine or ten, the sight of a stranger may be seen to arouse an unmistakable fear response. Some discussion of this response is already given in Volume I (Chapter 15). Since then further study of the genesis of the response has been undertaken by Bronson and by Scarr & Salapatek in the United States and by Schaffer in the United Kingdom. When account is taken of the different experimental situations and different methods of scoring responses, findings are highly compatible.
85
Fear of Strangers
Using videotape-recording and sensitive measures of response, Bronson (in press) has studied reactions to strange persons in thirty-two infants in their familiar home surroundings, during their development from the age of three to nine months. He reports that, usually starting at about four months, most infants will occasionally respond to a strange person with a cry, a whimper, or a frown, and that these wary responses start to appear at the age when undiscriminating smiling at strangers begins to wane. Throughout the fourth and fifth months, however, visual discrimination of strangers remains slow and uncertain. An infant of this age may spend long periods staring intently at a nearby stranger and may delay a long time before responding; and, on occasion, his response may change from a smile to a frown. Whether wariness is shown is determined by such
-100-
variables as the visual characteristics of the stranger, his proximity and manner of approach; but before six months, in contrast to later (see below, pp. 120 -2), it makes little difference whether or not an infant is held by, or can see, mother. At this age, moreover, the response in any one infant is far from stable.
Once infants are past six months, these responses usually become more differentiated and, for any one infant, more predictable. In the first place the response is more clearly aversive, so that the term fear becomes more applicable. In the second, perceptual identification of a stranger is posing fewer difficulties. Even so, as Schaffer ( 1971) points out, the earliest occasions on which an infant shows fear of strangers are when mother is herself present and the infant, by looking to and fro, can compare the two figures. Only later is an infant able readily to make the comparison, from memory, when his mother is absent.
As an infant reaches the end of his first year, his responses become still more predictable, and he may show unease either towards a particular person or towards members of a particular sex.
A point emphasized in the first volume is that, in any one infant, fear of strangers varies greatly according to conditions. How far distant from the infant a stranger is, whether he approaches, and whether he touches the infant are all extremely important; so also is the distance the child is from his mother. The significance of these variables for an understanding of fear is discussed later in this chapter and in those that follow.
Fear of Strange Objects
At about the same age as a child is becoming afraid of strangers he is apt also to become afraid of the sight of novel situations and strange objects. For example, Meili ( 1959) found, during a longitudinal study, that many infants become afraid of a jack-in-the-box at about ten months. Support for this finding comes from Scarr & Salapatek ( 1970) who undertook a crosssectional study of fear responses as they occur in children between the ages of about five and eighteen months. At every age between nine and fourteen months rather more than a third of the children were frightened both by the jack-in-the-box test and by the approach of a mechanical dog. Fewer of the younger and of the older children were afraid of these situations.
86
Schaffer has investigated the development of responses to unfamiliar objects. In a series of experiments, Schaffer & Parry
-101-
( 1969; 1970) have shown that at the age of six months, although infants are fully capable of perceiving differences between a familiar object and an unfamiliar one, they none the less approach both kinds of object without any discrimination. From about eight months of age onwards, however, infants begin to show sharp discrimination. Thenceforward, whereas a familiar object is approached with confidence, an unfamiliar one is treated with caution: in the experiments some children merely stared at the unfamiliar object, others appeared as if frozen, while yet others were distressed and withdrew. Even when, after they had become familiar with the object, they began to touch it, they still did so only briefly and tentatively.
A particularly interesting observation reported by Schaffer ( 1971) is the way in which a one- year-old infant habitually turns to his mother when uncertain, whereas an infant aged six months does not. Two groups of infants, one aged six months and the other twelve months, were presented with an array of stimulus objects. Behind each infant sat his mother, with instructions to say and do nothing unless her infant became upset. Whereas the younger infants seemed to be spellbound by the objects in front of them and unaware that mother was immediately behind, the older infants turned frequently from the objects to mother and back again, apparently well able to keep mother in mind despite her being perceptually absent. Thus by twelve months of age an infant is capable of organized fear behaviour characterized typically by movement away from objects of one class and towards objects of another. The development in an infant, during the second half of the first year, of the capacity to turn to mother when he is frightened and to find comfort in her presence is described in greater detail in the final section of this chapter.
Certain other conditions that regularly arouse fear behaviour in human infants during the second half of the first year are constellations of visual stimuli that act as naturally occurring clues to the imminence of two dangers common in the wild: the danger of falling, and the danger of being attacked or overwhelmed by an object rapidly approaching.
Fear of the Visual Cliff
Walk & Gibson ( 1961) have described the behaviour of thirtysix infants, aged from six to fourteen months, all of whom could crawl, when tested on an apparatus known as the 'visual
-102-
cliff'. This consists of a board laid across a sheet of heavy glass, with a patterned material that is directly beneath the glass on one side and, by dropping vertically, is several feet below it on the other. The infant is placed on the centre of the board while his mother stands at one or other side, calling her child to come to her across the glass-covered table which, according to the side at which she is standing, appears to be either solid table or deep chasm. Since each mother alternated her position it was easy to determine whether or not an infant was afraid to venture across the glass-covered 'chasm'.
Of the thirty-six infants tested only three, all boys, crossed the chasm to get to mother. All the others refused: some cried, others backed away to avoid the chasm, others peered through the
87
glass or patted it. Yet, when mother was on the 'solid' side, most of the infants crawled quickly towards her. In the majority, then, discrimination was very evident.
Scarr & Salapatek ( 1970) repeated the experiment with their sample and found that the older the child the more likely was he to refuse to cross the chasm. Whereas of the children aged between seven and eleven months nearly half were willing to cross to get to mother, all those aged thirteen months and over refused to do so.
Since Walk & Gibson tested the young of many animal species on the visual cliff, it was possible for them to draw broad conclusions. Certainly in other species, and probably in humans as well, it is clear that fear on perception of clues indicative of height develops very early, and even when an infant has had no experience of falling. The perceptual cue that appears to trigger off avoidance behaviour is 'motion perspective', namely the differential motion of foreground and background produced by the infant's own actions. Compared with lambs and kids, all of which show reliable discrimination and accurate avoiding movements from the beginning, human infants are both less reliable in discriminating and more clumsy in movement. Nevertheless, a strong bias to avoid the chasm was evident in all but a small minority.
Fear of an Approaching Object (Looming)
Another stimulus condition that seems to elicit a natural fear reaction in human infants and to do so very early in life is a visual stimulus that expands rapidly, which is habitually interpreted by adults as indicating something rapidly approaching.
Many years ago Valentine ( 1930) noted that approach elicits
-103-
fear in young children. A small girl of fourteen months, he reported, showed great fear of a teddy bear whenever it was moved towards her but would pick it up and kiss it whenever it was still.
In recent times Boweret al. ( 1970) have shown that a defensive response to an approaching object occurs in infants as young as two weeks (provided they are alert in an upright or near upright position). In a study of over forty infants they report that, each time a soft object (foam-rubber cube 20 cm per side) approaches to within about eight inches of a baby's face, without touching him, the infant pulls his head backwards, puts his hands between his face and the object, and cries loudly. The closer the object comes, the louder the cry. Further tests show that, when the stimulus consists merely of a shadow rapidly expanding on a screen, the response is similar though less intense. By contrast, when an object moves away there is no response. In the next chapter it is seen that young rhesus monkeys behave in very similar ways.
The fear-eliciting properties of an approaching or looming object have probably been underestimated in the past; and it seems likely that, in some of the experiments on an infant's response to strangers and to novel objects, approach of the stranger or object has played a larger part in determining a fear response than the experimenters realized.
88
A stimulus condition related to looming and approach is darkness. During the first year of life, fear of darkness, which is common in later years, is not very evident. Even so, by the age of ten months, infants are more likely to leave mother to enter and explore a brightly illuminated room than a dimly lit one ( Rheingold & Eckerman 1970).
Fear of an Anticipated Situation
Yet another situation that arouses fear, observable towards the end of the first year but not earlier, is when a baby uses current clues to anticipate something unpleasant. Levy ( 1951) describes the behaviour of babies of different ages when they catch sight of a doctor preparing to repeat an injection first given a few weeks earlier. Before eleven months of age only a very occasional infant was observed to react with fear. At ages eleven and twelve months, however, one-quarter of the sample did so. In such cases, it seems probable, learning from experience has occurred.
-104-
Thus by the end of the first year an infant is withdrawing in an organized way when he perceives any of a number of stimulus situations that can be regarded as naturally occurring clues to potentially dangerous situations. Further, he has learnt a good deal about his perceptual world. Towards the familiar and the strange, and towards what he has learnt is agreeable and what disagreeable, he is behaving with a rough-and-ready discrimination. He moves towards the one and away from the other.
Fear-arousing situations: the second and later years Sources of Data
It has already been noted how few researchers have made systematic attempts to study the situations that commonly arouse fear in humans. Most of the few data published during recent decades come from various longitudinal studies of children developing. Examples are a study by Macfarlane, Allen & Honzik ( 1954) of about one hundred children in California and another by Newson & Newson ( 1968) of 700 children and their parents in an English urban community. In none of these studies, however, was the nature of the situations that arouse fear at the centre of interest, nor was the information reported obtained either from direct observation or from interview of the children themselves. The latter limitation applies also to the findings of a cross-sectional study of nearly 500 children undertaken by Lapouse & Monk ( 1959) in New York State. 1 In all these projects information came only from mothers.
Because of the paucity of recently gathered data it is necessary to turn to the results of work done during the early years of child development research.
Some forty years ago an American psychologist, A. T. Jersild, began a series of studies in which he set out to describe the kinds of situation in which children exhibit fear, and how these change as a child grows older. 2 Different methods of obtaining information were employed in different studies. The four main methods were: day-to-day recording by parents; simple experiments; interviewing children about situations of which
____________________
1 A representative sample, distinct from the smaller sample referred to on p. 98 above.
2 Jersild's principal studies are published as monographs: Jersild Markey & Jersild ( 1933)
89
and Jersild & Holmes ( 1935a). Abstracts of these and other studies with full references are in the symposium on Child Behavior and Development edited by Barker, Kounin & Wright ( 1943), and also in Jersild ( 1947)
-105-
they were currently afraid; giving questionnaires to adults about what they recalled of the situations that made them afraid as children. The subjects of each study were different and were drawn from different age-ranges. Despite a number of shortcomings, these studies are still much the most extensive yet attempted and therefore remain principal sources. Not only do the findings conform with ordinary experience, but at a number of points they are supported and amplified both by earlier (e. g. Hagman 1932) and by more recent work.
Findings from Parents' Records and Naturalistic Observations
The aim of one of Jersild's studies was to obtain a detailed record of the occasions and situations in which ordinary children exhibit fear during the course of their everyday lives. To this end the parents of over one hundred young children were enlisted, all of whom were prepared to keep detailed records of every occasion on which their child showed fear during a period lasting twenty-one days. Mimeographed forms and instructions were issued. On each occasion that a child showed fear parents were asked to record: (a) the actual behaviour exhibited (e. g. startle, withdrawing, turning to an adult, cries or other vocalizations, and words spoken); (b) the situation in which the behaviour occurred, in terms not only of its apparent cause (specific stimulus) but of the setting (place, time, what child was doing, persons present); and (c) the child's current condition (well or ill, fresh or tired).
In all, 136 records were obtained of children between twelve and fifty-nine months of age (the records of a few younger and older children were too few to give useful results). The sample was biased towards the higher end of the socio-economic scale. A majority lived in a large city, but there were also children living in suburbs, small towns, and rural areas. Distribution by age was as follows: second year, twenty-three children; third year, forty-five; fourth year, forty-six; and fifth year, twenty-two children.
The authors were struck by how few were the occurrences of fear recorded by the parents of children in these age-groups, a finding confirmed when observations were made of some of the same children in a nursery school. The occurrences recorded during the three weeks for children in the two younger agegroups averaged only six per child, or two a week. For children in the two older age-groups the average was three and a half per child, or barely more than one a week. For about one in
-106-
ten of the children at each age-level no occurrence of fear was recorded throughout the three- week period. Although these figures suggest that some occurrences of slight or transient fear may have gone unrecorded, there is independent evidence that at least some children during their second year show fear very infrequently. For example, Valentine ( 1930), who kept daily notes on his children, was impressed by the infrequency with which he saw fear responses in them and describes how surprised he was when a child who had fallen and hurt himself would none the less immediately go climbing again. Anderson ( 1972a), who observed fifty-two children between the ages of twelve months and three years two months in a London park,
90
also remarks how infrequently the children exhibited fear. Occurrences, he reports, were 'uncommon and short-lived'. It should, however, be noted that in both these studies and also in Jersild's the children observed were not alone. The difference made by presence or absence of a trusted adult cannot be exaggerated (see the final section of this chapter). When we turn to examine the kinds of situation that are reported to arouse fear we find remarkably little change between the second and the fifth years. From the mother's day-to-day records reported by Jersild six situations stand out as being likely to arouse fear at least occasionally in a fairly large proportion of the children at each age-level:
--noise, and events associated with noise
--heights
--strange people or familiar people in strange guise --strange objects and surroundings
--animals
--pain or persons associated with pain.
In each of these six situations some 40 per cent of all the children were recorded as having shown fear behaviour on some occasion during the three-week spell. In so far as there was any reduction with age in the proportion of children showing fear the reduction occurred after the third birthday. 1
____________________
1 When analysed by age the picture from these six situations is as follows. Of the one-year-
olds, 60 per cent were recorded as having shown fear of noise, 52 per cent fear of pain or potential pain, and from 35 to 40 per cent as having shown fear in each of the other four situations, of which one was the presence of animals. Of the four-year-olds, only 23 per cent were recorded as having shown fear of noise and events
-107-
Among the many other situations recorded as having aroused fear, but in a smaller percentage of these children, were sudden unexpected movements, especially when an object was both approaching and noisy, and also bright lights, flashes, etc. Together, these kinds of situation elicited fear in nearly 30 per cent of the one- and two-year-olds; but of the older children no more than about 10 per cent had shown fear in those situations. The dark, especially being alone in the dark, was recorded as having aroused fear in about 10 per cent of the children during the three-week period of observation: here there was no change with age. Fear of being left alone or abandoned was recorded for about 10 per cent of the children at each of the age- levels. Only after the second birthday were any of the children recorded as having shown fear of imaginary creatures, the proportion being about 6 per cent. The origin and nature of such fears are discussed in Chapters 10 and 11.
The forms of behaviour recorded as shown by these children when afraid differed very little between the younger and the older ones. The behaviour most frequently recorded was crying in its various forms, from whimpering to screaming and including explicit cries for help. In each age-group no less than onethird of the episodes of fear recorded by mothers were signalled by crying of one sort or another. Another form of behaviour also frequently recorded was either turning towards an adult or running to him or her, with or without clinging: in each age-group about one-sixth of the fear-producing situations were said to elicit behaviour of that sort. Avoiding action or running away was recorded in about one-fifth of the episodes. In the remainder, fear was inferred because a child trembled or jumped, showed a frightened
91
expression, hid his head, or stayed unwontedly still. Occasionally a child was aggressive, or protective of another.
In this catalogue of forms of behaviour held by mothers to be indicative of fear it should be noted that two of the commonest are crying for or turning towards a protective figure. This finding resembles that of Anderson ( 1972a) when he interviewed the mothers of eighteen two-year-olds in London. The forms of fear behaviour most commonly described to him were screaming, crying, turning to mother, clinging to or
____________________
associated with noise, but no fewer than 40 per cent had shown fear of animals, the same percentage as for the one-year-olds. In each of the other situations, however, including the possibility of pain, only about 15 per cent of the four-year-olds had shown fear.
-108-
following mother, and staying close to her. Withdrawing from the frightening object was referred to less frequently.
When, however, Anderson was himself observing the fear behaviour of another but very similar sample of toddlers while they were with mother in a London park (see p. 107 above), rather different forms of fear behaviour were seen. In the dozen episodes observed, the fear- arousing object was an approaching animal (eight cases), an approaching child (three cases), and noise (one case). In such conditions a toddler would suddenly cease his activity, back away from whatever it was that was frightening him while continuing to fixate it, and simultaneously edge towards his mother. Crying was not observed. Once the object had moved away, the child would advance again but continue to fixate it.
The differences in forms of behaviour exhibited turn presumably on the intensity of fear. At high intensity crying and clinging are common; at low intensity withdrawing from the object and backing towards mother.
A limitation of records kept by parents and also of naturalistic observations such as those of Anderson is that, when a child is not described as having shown fear of a particular class of situation during the period his behaviour was recorded, it remains uncertain whether he is never afraid in such situations or whether it just happened that during that span of time he was never confronted by one when his mother or the observer was about. The experiments devised by Jersild & Holmes help to clarify this matter, though they too have obvious limitations.
Experimental Findings
It is evident that ethical considerations limit severely the kinds of experiment it is legitimate to do in order to explore the situations that arouse fear in humans, especially in young children.
In their experimental work with children between their second and sixth birthdays, therefore, Jersild & Holmes took many precautions. In the first place, each child was throughout with an adult who was experienced with children, and who had made a friendly contact with him before the experiments began. In the second, each situation with which a child was confronted
92
was one that many children hardly find frightening. In the third, a child was introduced to the situation by easy stages. And finally, if a child refused to take part, the experiment was ended.
-109-
There were eight potentially fear-provoking situations. Four were presented on the first day, extending over about fifteen minutes; two on the day following; and two about four weeks later. For a few minutes between each fear-inducing situation a child was allowed to play with toys. Since it appears that the situations were presented to every child in the same order, there is a serious possibility that responses to later situations may have been influenced by the experience of earlier ones; though in which direction such influence may have been exerted it is difficult to know. On the one hand is the possibility that, through habituation, responses to later situations may have been less than they would otherwise have been. On the other is the possibility that, as the series progressed, a child may have become increasingly sensitized and therefore have shown more fear in some of the later situations than he would otherwise have done. The finding that a larger proportion of children showed fear in the later experimental situations than in the earlier ones is in keeping with the latter possibility. The eight situations were chosen because earlier studies had shown that they were likely to elicit at least a little fear in a substantial proportion of young children. Details are as follows:
1. Being left alone: When the child is seated at a table playing with a toy, the experimenter names a pretext for leaving the room (which until the time of the experiment was unfamiliar to the child). The experimenter remains outside the room for two minutes. The child's behaviour is recorded by concealed observers.
2. Sudden displacement or loss of support: A bridge-like piece of apparatus, consisting of two boards laid end to end at a height of about two inches above the floor, was used. The first board is securely supported, but when the child steps onto the second board, which is supported only at the middle, it gives way and descends to the floor.
3. Dark passage: While playing ball with the child, the experimenter seemingly inadvertently throws the ball into a dark passage, 18 ft long, leading from one corner of the room. The child is asked to retrieve the ball.
4. Strange person: While the child is temporarily withdrawn from the room an assistant, dressed in a long grey coat, a large black hat, and a veil that obscures her features, seats herself in one of the two chairs near the entrance. The child
-110-
returns and his reactions are observed when he notices the stranger and when he is asked to obtain toys placed near the stranger's chair.
5. High board: A board 12 in. wide, about 8 ft long, and 2 in. thick, held firmly in place at the ends by two stationary holders, is arranged at various heights from the ground and the child is asked to walk from one end of the board to the other to obtain a box of brightly coloured toys. The board is first placed at a distance of four feet from the floor; it is subsequently lowered if the child refuses to walk across at this height and is raised if he performs at the four-foot level.
6. Loud sound: An iron pipe 2 ft long and 2 1/4 in. diameter, suspended from the ceiling in a corner of the room behind a screen, is struck a sharp blow with a hammer while the child and the experimenter are seated at a table containing toys. The child's response to the unexpected noise from an unseen source is first observed and then, pointing to the screen, the experimenter asks the child to 'Go and see what made that noise'.
7. Snake: A snake, harmless and about 2 ft long, was placed in a box sufficiently deep to 93
ensure that it could not immediately climb out when the top was removed. In the box was placed a small coloured toy. The child's attention is directed to the box, the lid is uncovered, and the child is allowed to look in; if he raises any questions the experimenter simply says 'It is a snake', and then points to the toy and asks the child to reach in and get the toy.
8. Large dog: While the child is seated at a table with toys a large collie dog is brought into the room on a leash by a familiar person. The dog is led to a certain point in the room and, after preliminary comment by the experimenter, the child is asked to go and pat the dog.
The subjects were 105 children, half from a private nursery school for better-off families and half from a public nursery school for poorer families. Fifty-seven were boys and fortyeight girls. Children were tested only when they were fit, willing, and in a good humour; and the experiments were never combined with any other sort of examination. On each occasion when tests were carried out, two-year-olds and three-yearolds were well represented (never fewer than twenty-one children tested and usually between thirty and forty-five); but
-111-
four-year-olds and five-year-olds were rather few (numbers tested varied between seven and fourteen children). Each experiment, except the first, was presented to a child in four stages: first he was given directions what to do; then, if he was hesitant, he was given reassurance and encouragement; then, if he still refrained, the experimenter offered to accompany him in the task; finally, if a child was still reluctant to take part, the experiment was abandoned.
