Whereas in both Jensen's and Harlow's
laboratories
the infants had been brought up with mother alone, each pair in a small cage, in the laboratories of Hinde
-63-
and of Kaufman infant and mother lived as part of a stable social group in a fairly large cage.
-63-
and of Kaufman infant and mother lived as part of a stable social group in a fairly large cage.
Bowlby - Separation
In some studies and at some ages no differences are observed in the behaviour of boys and girls.
In so far as any differences are observed, boys tend to explore more in mother's presence and to be more vigorous in their attempts to reach her when she has gone; girls tend to keep closer to mother and also to make friends more readily with the stranger.
A further finding from these miniature separation experiments, and one that links with the findings of Shirley ( 1942) and Heathers ( 1954) (see pp. 35 and 37 above), is only very recently reported. This is that when a child of about one year is tested in Ainsworth's series of episodes for a second time, a few weeks after the first testing, he is more upset and anxious than he was on the first occasion. When mother is present he keeps closer to her and clings more tightly. When she is absent he cries more ( Ainsworth, personal communication). These findings emerge from a test--retest study of twenty-four babies tested first at fifty weeks of age and a second time two weeks later. On the assumption that increased sensitivity is not due simply to maturation, which is unlikely, these findings provide the first experimental evidence that at one year of age a separation lasting only a few minutes, in what would ordinarily be regarded as a bland situation, is apt to leave a child more sensitive than he was before to a repetition of the experience.
Ontogeny of responses to separation The First Year
Since the responses to separation that are so unmistakable in infants of twelve months and older are not present at birth, it is clear that they must develop at some time during the first year of life. Unfortunately, studies designed to throw light on this development are few, and are confined to infants admitted to hospital. Nevertheless such evidence as is available is unambiguous. It is in keeping, moreover, with what is known about the
-52-
development of attachment behaviour and about cognitive development generally.
In Chapter 15 of the earlier volume the steps by which, during the early months of life, an infant's attachment behaviour gradually becomes focused on a discriminated and preferred attachment figure are described. Development can be summarized as follows: before sixteen weeks differentially directed responses are few in number and are seen only when methods of observation are sensitive; between sixteen and twenty-six weeks differentially directed responses are both more numerous and more apparent; and in the great majority of family infants of six months and over they are plain for all to see. It comes as no surprise, therefore, that the full range of responses to separation described in earlier sections of this chapter is not seen before six or seven months of age.
Schaffer studied seventy-six infants of various ages under twelve months admitted to hospital: none was marasmic, deformed, or thought to be brain-damaged. Of the total, twentyfive were
51
healthy infants admitted for elective surgery. While in hospital each child was observed during a two-hour session on each of the first three days (see Schaffer 1958; Schaffer & Callender 1959). Infants were not only without mother but had very little social interaction with nurses.
The responses observed in these twenty-five healthy infants differed greatly according to the child's age. The dividingpoint was twenty-eight weeks. Of the sixteen aged twenty-nine weeks and over, all but one fretted piteously, exhibiting all the struggling, restlessness, and crying so typical of two- and threeyear-olds. Of the nine aged twenty-eight weeks and under, by contrast, all but two 1 are reported to have accepted the situation without protest or fretting: only an unwonted and bewildered silence indicated their awareness of change.
Schaffer emphasizes that the shift from a bewildered response to active protest and fretting occurs suddenly and at full intensity at about twenty-eight weeks af age. Thus, of the sixteen infants aged between twenty-nine and fifty-one weeks, both the length of the period of fretting and the intensity of it were as great in those of seven and eight months as in those of eleven and twelve months.
Furthermore, responses both to the observers and to mother when she visited changed equally suddenly at about thirty weeks:
____________________
1 One of the exceptions was an infant already twenty-eight weeks of age.
-53-
In the younger infants [twenty-eight weeks and less] most observation sessions showed these infants to be normally responsive, although the people confronting them were complete strangers. This held as much for the nurses who fed and bathed them as for the observers. . . . In the older group, on the other hand, normal [i. e. friendly] responsiveness was almost completely lacking, and the majority of observation sessions showed these infants to be negative and frightened when approached by a stranger-- a type of behaviour not seen at all in the younger group ( Schaffer & Callender 1959).
Although there were too few observations for statistical comparisons to be made of how infants of different ages responded to mother's visit, such observations as could be made support the thesis of a sharp change around twenty-eight to thirty weeks. Infants older than that mostly clung rather desperately to mother, behaviour that was in striking contrast to their negative responses to the observers. The younger infants, by contrast, tended to respond to mother and to observers without showing marked discrimination between them. Similarly, when mother departed, whereas older infants cried loudly and for a long time, even desperately, the younger ones showed no sign of protest.
Finally, the behaviour of the infants on their return home from hospital differed greatly according to age-group. Most of the infants aged seven months and over showed intense attachment behaviour. They clung almost continuously to mother, cried loud and long when left alone by her, and were notably afraid of strangers. Even figures formerly familiar, such as father and siblings, were sometimes regarded with suspicion. Infants aged under seven months, by contrast, showed little or no attachment behaviour during their early days at home. Their mothers described them as 'strange'. On the one hand, these infants seemed utterly
52
preoccupied in scanning the environment; on the other, they seemed unheedful of adults or perhaps averted their head when approached:
For hours on end sometimes the infant would crane his neck, scanning his surroundings without apparently focusing on any particular feature and letting his eyes sweep over all objects without attending to any particular one. A completely blank expression was usually observed on his face, though sometimes a bewildered or frightened look was reported. In the extreme form of this syndrome the infants were quite inactive throughout, apart from the scanning behaviour, and
-54-
no vocalization was heard though one or two were reported to have cried or whimpered. When confronted with a toy the infant disregarded it.
To attempts by adults to make contact with them some of these younger infants seemed altogether oblivious. Others seemed to avoid the adults, and others again to gaze 'through' them with the same blank look that they used for the rest of the environment.
The only way in which the responses of the infants of the two age-groups were similar was in regard to sleep: in infants of both groups disturbed sleep and night-crying were common.
How the responses of infants of under seven months are best understood, and what their significance for an infant's future development may be, is difficult to know. It is plain, however, that the responses of these younger infants to separation are very different at every phase from those of older ones, and that it is only after about seven months of age that the patterns that are the subject of this work are to be seen.
In discussion of his findings Schaffer ( 1958) draws on Piaget's work on the development of an infant's concept of an object ( Piaget 1937). Only during the second half of the first year, Piaget finds, is there evidence that an infant is beginning to be able to conceive of an object as something that exists independently of himself, in a context of spatial and causal relations, even when it is not present to his perception, and so to search for it when it is missing. Bell ( 1970) confirms Piaget's findings and, in addition, reports the results of an experiment designed to test whether or not an infant develops a capacity to conceive of a person as a persisting object earlier than he develops the capacity in regard to inanimate things. Although her results show that a majority of infants develop the capacity in regard to a person earlier than in regard to things, it is not until about the ninth month that the capacity in regard to persons is reasonably well developed, and in a minority it lags some weeks behind that. For reasons to do with cognitive development, therefore, the types of response to separation with which we are concerned could hardly be expected in infants younger than those in whom they are seen.
Change after the First Birthday
All the evidence suggests that, once established, the typical patterns of response to being placed in strange surroundings
-55-
53
with strange people do not undergo marked change, either in form or in intensity, much before the third birthday. Thenceforward their intensity begins to diminish, but only slowly. As an example, the change in Lottie's feelings about going to nursery school that appeared at the beginning of the third term, when she was aged two years and nine months (see above, p. 38 ), is characteristic of many children. Provided he knows where his mother is and has good reason to expect her to return soon, a child begins to accept another fairly familiar person, even when he is in a fairly unfamiliar place.
The only conditions at present known that reduce appreciably the effects of separation from mother are familiar possessions, the companionship of another and familiar child and, as Robertson & Robertson ( 1971) have shown, especially mothering from a skilled and familiar foster mother. By contrast, strange people, strange places, and strange proceedings are always alarming; and they are especially alarming when encountered alone (see Chapters 7 and 8). ?
Since distress at being separated unwillingly from an attachment figure is an indissoluble part of being attached to someone, changes occurring with age in the form of response to separation accompany, step by step, changes in the form that attachment behaviour takes. These changes are sketched in the first volume (Chapters 11 and 17) and need not be described further here. In so far as attachments to loved figures are an integral part of our lives, a potential to feel distress on separation from them and anxiety at the prospect of separation is so also. That is a theme running through the rest of this volume.
Meanwhile, in order that we may view the responses to separation seen in humans in a perspective broader than has been traditional, it is useful to compare the responses of young human children with those of the young of other species. When that is done it becomes evident that, just as attachment behaviour occurs in rather similar forms across a number of mammalian and avian species, so also do responses to separation. Here again man is no isolated case.
-56-
54
Chapter 4
Behaviour with and without Mother: Non-human Primates
Man with all his noble qualities, with sympathy that feels for the most debased, with benevolence which extends not only to other men but to the humblest living creature, with his god-like intellect which has penetrated into the movements and constitution of the solar system -- with all these exalted powers -- still bears in his bodily frame the indelible stamp of his lowly origin.
CHARLES DARWIN ( 1871)
Naturalistic observations
It has long been known that isolation and separation from a mother figure can cause distress, expressed in calling and searching, in the young of many species of bird and mammal. The 'lost piping' of young ducklings who have become attached to and have temporarily lost a mother figure is a familiar example. Others are the bleating of lambs and the yelping of puppies. Coming nearer man, there are numerous examples in the accounts of monkey and ape infants brought up by human caretakers. All accounts agree on the intensity of protest exhibited whenever a baby primate loses its mother figure, and the intensity of distress that follows when she cannot be found. All agree, too, on the intensity of clinging that occurs after the two are reunited.
For example, Bolwig ( 1963), in his account of the little patas monkey he reared from a few days old,1 describes how from the first the little monkey 'showed no fear of man, cried much, and panicked when left alone. . . . The screaming, with wide open mouth and distorted face, was only heard when the observer moved out of the monkey's immediate reach or sight. On such occasions he would more often than not run staggering to the nearest person in sight. ' Soon the monkey's attachment had become focused on Bolwig himself and then, until three and a
____________________
1 See Volume I, Chapter 11.
-57-
half months of age, the monkey could be very troublesome unless perpetually with his caretaker.
By the age of four months, however, the little monkey was exploring increasingly far afield and his master decided
to leave him for some hours every day in a cage with other monkeys of his own kind. This attempt was, however, not very successful. Although he knew the other monkeys well and was accustomed to play with them he panicked as soon as he knew I wanted to leave him
55
behind, screamed, clung desperately to me and then tried to tear the door open. He would sit and cry until I finally let him out. Afterwards he would cling to me and refuse to leave me out of sight for the rest of the day. In the evening when asleep he would wake up with small shrieks and cling to me, showing all signs of terror when I tried to release his grip.
Accounts of similar behaviour are given of infant chimpanzees. Cathy Hayes ( 1951) recounts how Viki, a female she adopted at three days, would, when aged four months, cling to her foster mother
from the moment she left her crib until she was tucked in at night. . . . She sat on my lap while I ate or studied. She straddled my hip as I cooked. If she were on the floor, and I started to get away, she screamed and clung to my leg until I picked her up. . . . If some rare lack of vigilance on her part let a room's length separate us, she came charging across the abyss, screaming at the height of her considerable ability.
The Kelloggs, who did not adopt their female chimp, Gua, until she was seven months old and who kept her for nine months, report identical behaviour ( Kellogg & Kellogg 1933). They describe
an intensive and tenacious impulse to remain within sight and call of some friend, guardian, or protector. Throughout the entire nine months . . . whether indoors or out, she almost never roamed very far from someone she knew. To shut her up in a room by herself, or to walk away faster than she could run, and to leave her behind, proved, as well as we could judge, to be the most awful punishment that could possibly be inflicted. She could not be alone apparently without suffering.
Comparing Gua with their son, who was two and a half months older than she, the Kelloggs report:
-58-
Both subjects displayed what might be called anxious behaviour (i. e. fretting and crying) if obvious preparations were being made by the grown-ups to leave the house. This led (in Gua) to an early understanding of the mechanism of door closing and a keen and continual observation of the doors in her vicinity. If she happened to be on one side of a doorway, and her friends on the other, the slightest movement of the door toward closing, whether produced by human hands or by the wind, would bring Gua rushing through the narrowing aperture, crying as she came.
The very detailed observations made by van Lawick-Goodall ( 1968) of chimpanzees in the Gombe Stream Reserve in central Africa show not only that anxious and distressed behaviour on being separated, as reported of animals in captivity, occurs also in the wild but that distress at separation continues throughout chimpanzee childhood. During the first year an infant is rarely out of actual contact with mother and, although from its first birthday onwards it spends more time out of contact, it none the less remains in proximity to her. Not until young are four and a half years of age are any of them seen travelling not in the company of mother, and then only rarely. 1
56
Once an infant begins to spend time out of contact with mother, proximity is maintained largely by auditory signals. Either mother or infant emits a 'hoo' whimper to which the other promptly responds:
When the infant . . . begins to move from its mother, it invariably utters this sound if it gets into any difficulty and cannot quickly return to her. Until the infant's locomotor patterns are fairly well developed the mother normally responds by going to fetch it at once. The same sound is used by the mother when she reaches to remove her infant from some potentially dangerous situation or even, on occasion, as she gestures it to cling on when she is ready to go. The 'hoo' whimper therefore serves as a fairly specific signal in re-establishing mother-- infant contact.
Another signal used by infants is a scream; it is elicited whenever an infant falls or nearly falls from its mother or is frightened by a sudden loud noise. When her infant screams a mother almost unfailingly retrieves it and cradles it: 'On several occasions
____________________
1 For a brief description of the developmental course of attachment behaviour in
chimpanzees see Volume I. Chanter 11. -59-
infants screamed loudly when their mothers started to move away without them. Each time the mothers immediately turned back and retrieved them. Indeed, throughout infancy, screaming normally results in the mother hurrying to rescue her child. '
Juveniles up to five or six years old also scream when lost or in trouble, and again mother usually hurries to the rescue:
On several occasions juveniles were observed who had accidentally lost their mothers. In each instance, after peering round from various trees, whimpering and screaming as they did so, they hurried off -- often in the wrong direction. On three occasions I was able to observe the reaction of the mother and every time, although she set off in the direction of her offspring's screaming, she herself made no sound to indicate her whereabouts.
In one case a juvenile female aged five years lost her mother in the evening and was still whimpering and crying the following morning. In another case, a juvenile stopped screaming before her mother found her, which resulted in a separation lasting several hours. (No information is available regarding behaviour of the young after reunion. )
Thus in these wild-living chimpanzees proximity of young to mother is maintained until pre- adolescence. Separations are rare, and usually quickly rectified by vocal signals and mutual search.
Early experimental studies
These naturalistic accounts show plainly not only that the attachment behaviour of young non- human primates is very similar to the attachment behaviour of young children but that their responses to separation are very similar also. Because of this, and because experimental separations lasting longer than minutes are inadmissible in the case of human young, more
57
than one scientist has turned to monkey young for experimental subjects. A number of studies from at least four different centres are now published. Animals used include infants aged between two and eight months, of five different species, namely four species of macaque (the rhesus, the pigtail, the bonnet, and the Java) and the patas monkey. All five are species of semiterrestrial and group-living old-world monkeys. 1
____________________
1 The nearest experiments so far reported with chimpanzees are those by Mason ( 1965), but
in this case the separations were from a cage-mate of the same age and not from mother. -60-
Responses to separation differ as between species, though differences are more of intensity than of kind. In the case of rhesus, pigtail, and Java macaques great distress is observed throughout the period of separation itself and, afterwards, there is a very marked tendency to cling to mother and to resist any attempt at a further separation, however brief. In the case of both bonnet macaques and patas monkeys, intense distress is again seen during the first hours after separation, but then it wanes; thereafter activity is less depressed than in the other species of macaque and there is much less disturbance after reunion with mother. The reduction of distress in the bonnet macaques appears to come about in great part because the separated infant receives continuous substitute care from one of the other familiar females in the group.
In what follows attention is given to the studies using rhesus and pigtail infants both because their responses appear to resemble more closely those of human infants and because the studies of these species are more numerous and extensive, especially in the case of the rhesus. Those wishing to compare the behaviour of bonnet macaques are referred to the study by Rosenblum & Kaufman ( 1968; see also Kaufman & Rosenblum 1969); and of patas monkeys to the study by Preston, Baker & Seay ( 1970). A useful review of separation studies is given by Mitchell ( 1970).
An early experimental study was carried out by Jensen & Tolman ( 1962). When two infant pigtail monkeys, each reared in a cage alone with mother, were aged respectively five and seven months, the infants and mothers were exchanged on several occasions for periods of no longer than five minutes. Observation was through a one-way vision screen.
Because mother and infant cling tightly to one another separation cannot be achieved with monkeys except by deception or by the exercise of a good deal of force. The protests of both parties are intense. Jensen & Tolman give a vivid account:
Separation of mother and infant monkeys is an extremely stressful event for both mother and infant as well as for the attendants and for all other monkeys within sight or earshot of the experience. The mother becomes ferocious toward attendants and extremely protective of her infant. The infant's screams can be heard over almost the entire building. The mother struggles and attacks the separators. The baby clings tightly to the mother and to any object which it can
-61-
58
grasp to avoid being held or removed by the attendant. With the baby gone, the mother paces the cage almost constantly, charges the cage occasionally, bites at it, and makes continual attempts to escape. She also lets out occasional mooing-like sounds. The infant emits high pitched shrill screams intermittently and almost continuously for the period of separation.
As soon as the five minutes had elapsed and mother and infant were reunited each immediately went to the other and the two remained in the closest possible contact: 'The mother sits quietly holding her baby, and if no attendants are present she very quickly seems content and relaxed. All is quiet in the room. No more piercing screeches of the baby or sounds from the mother are heard. ' The duration of this unbroken period of intense mutual clinging following a separation that had lasted a mere five minutes was never less than fifteen minutes and in some cases as long as forty.
Other workers have subjected their monkey infants to much longer separations, the periods ranging from six days to as long as four weeks. In the case of pigtail and rhesus infants all observers report extreme and noisy distress during the twentyfour hours or so immediately after separation followed by a quieter period of a week or more during which the infants show little activity or play and, instead, sit hunched up and depressed.
Harlow has been responsible for two such studies. In one ( Seay, Hansen & Harlow 1962), four rhesus infants, ages ranging from twenty-four to thirty weeks, were kept apart from mother for a period of three weeks. 1 Since mother was in an adjacent cage and only a transparent screen separated the two, each could see and hear the other. Observations were made at regular intervals during the three weeks prior to separation, during the three weeks of separation, and for three weeks following separation. On each occasion two infants, already familiar with each other, were separated simultaneously and, during the period of separation, each infant had free access to the other. Thus throughout the period of separation all four infants had companionship, access to food and water,
____________________
1 For an account of attachment behaviour in rhesus monkeys see Volume I, Chapter 11.
Until it is about three years of age a young rhesus monkey in the wild remains close to mother.
-62-
and also visual and auditory contact with mother. Only physical contact with mother was missing.
As soon as the transparent screen had been lowered all four infants engaged in 'violent and prolonged protest'. There was much high-pitched screeching and crying, they made numerous attempts to reach mother, including hurling themselves against the screen, and they also scampered in a disoriented way around the cage. Later, when quiet, the infants huddled against the screen in as close proximity to mother as they could get. Initially mothers barked and threatened the experimenter, but their responses were less intense and persistent than the infants'. Throughout the separation period the pairs of separated infants showed little interest in one another and little play, in contrast to the active play between them seen in the three weeks prior to separation and after it was over. In the days after mother and infant had been reunited there was a very marked increase in the incidence of infant clinging to mother and
59
keeping in contact with her compared with what had occurred during the days before separation.
In a second and similar experiment Seay & Harlow ( 1965) separated a further eight rhesus infants from mother when the infants were aged thirty weeks. On this occasion the separation lasted only two weeks and mother was removed completely; a second separated infant was again available for play, but this time for only half an hour each day. Results were as before, including on the first day 'disoriented running about, climbing, screeching and crying'; relatively little interest was shown in the companion infant. After the phase of protest (duration of which is not reported) the infants 'passed into a stage characterized by low activity, little or no play and occasional crying'. The writers express the belief that 'this second stage is behaviorally similar to that described as despair of children separated from their mothers'. Immediately after reunion once again there was a phase of strong mutual clinging by mother and infant.
In 1966 and 1967 reports were published of closely similar findings by two other groups of workers, Spencer- Booth & Hinde using rhesus macaques in Cambridge, Kaufman & Rosenblum using pigtail macaques in New York. These two studies have much in common and are more informative than the earlier ones.
Whereas in both Jensen's and Harlow's laboratories the infants had been brought up with mother alone, each pair in a small cage, in the laboratories of Hinde
-63-
and of Kaufman infant and mother lived as part of a stable social group in a fairly large cage. Present with them were an adult male, two or three other adult females, and often other young. In both laboratories separation was effected by removing mother from the cage. This meant that the infant remained behind in a completely familiar environment with a number of other familiar animals: the only change in its life was the absence of mother.
A second advantage of the Hinde and the Kaufman studies is that results are reported in much greater detail, both in regard to the course of behaviour during the week or more of separation and in regard to the behaviour of both partners during the period of months, and in the Hinde study of nearly two years, after the separation was over. These observations are especially valuable in giving information about subsequent effects of the experimental separation.
In the experiment of Kaufman & Rosenblum ( 1967) the subjects were four infant pigtail monkeys ranging in age from twenty-one to twenty-six weeks. In each case mother was removed from the cage for a period of four weeks. The behaviour seen during the separation is reported to have fallen into three phases, described as 'agitation, depression and recovery'. Whereas three infants went through all three phases, the fourth, the daughter of the dominant female of the group, showed comparatively little depression and spent much of her time with the other adult females of her group. The behaviour of the remaining three infants is described as follows:
During the first phase pacing, searching head movements, frequent trips to the door and windows, sporadic and shortlived bursts of erratic play, and brief movements toward other members of the group seemed constant. Cooing, the rather plaintive distress call of the young macaque, was frequent. There was an increased amount of self-directed behaviour, such as sucking of digits, and mouthing and handling of other parts of the body, including the
60
genitals. The reaction persisted throughout the first day, during which time the infant did not sleep.
After 24 to 36 hours the pattern in three infants changed strikingly. Each infant sat hunched over, almost rolled into a ball, with his head often down between his legs. Movement was rare except when the infant was actively displaced. The movement that did occur appeared to be in slow motion,
-64-
except at feeding time or in response to aggression. The infant rarely responded to social invitation or made a social gesture, and play behaviour virtually ceased. The infant appeared disinterested in and disengaged from the environment. Occasionally he would look up and coo.
After persisting unchanged for 5 or 6 days the depression gradually began to lift. The recovery started with a resumption of a more upright posture and a resurgence of interest in the inanimate environment. Slow tentative exploration appeared with increasing frequency. Gradually, the motherless infant also began to interact with his social environment primarily with his peers, and then he began to play once again. The depression continued, but in an abated form. Periods of depression alternated with periods of inanimate-object exploration and play. Movement increased in amount and tempo. Toward the end of the month the infant appeared alert and active a great deal of the time, yet he still did not behave like a typical infant of that age.
Throughout the three months following mother's return the behaviour of mother and infant was recorded. In all four cases significant changes were observed, similar in direction to those observed in earlier studies:
When the mother was reintroduced to the group another dramatic change occurred. There was a tremendous reassertion of the dyadic relationship with marked increases in various measures of closeness in all four pairs. Clinging by the infant, protective enclosure by the mother, and nipple contact all rose significantly in the month after the reunion as compared to the frequency of these actions in the month before separation. Even in the third month after the reunion this trend was evident. This significant rise in measures of dyadic closeness is particularly striking in view of the fact that ordinarily for the age periods involved these particular behaviours fall considerably.
The increased closeness was manifest in other ways as well. A measure of mother--infant physical separation that we have found valuable in our normative studies concerns departures (usually by the infant) to another level of the pen. The frequency of such departures during the month after the reunion fell to 20 per cent of the departures in the month before the separation. Furthermore, the mean duration of these departures fell from 60? 5 seconds to 34? 4 seconds.
-65-
Not only did infants depart less frequently and for shorter periods than before the separation had occurred but mothers were more tolerant than before of their infant's persistent proximity, only rarely discouraging it by rejection or withdrawal.
61
Compared with the four subjects of Kaufman's experiment, the four subjects described initially by Spencer-Booth and Hinde were not only of a different species (rhesus instead of pigtail infants) but were rather older (thirty to thirty-two weeks of age instead of twenty-one to twenty-six weeks); and the length of time for which mother was removed was much shorter -- only six days in place of four weeks. Nevertheless, the behaviour observed both during the separation and during the months afterwards closely resembled what Kaufman and his colleagues were seeing (see Hinde, Spencer- Booth & Bruce 1966; Spencer- Booth & Hinde 1967).
Since their first reports Spencer-Booth and Hinde have published findings from a series of further studies in which the initial sample of four infants was increased to twenty-one, infants in certain sub-samples were exposed to a second separation, and another six infants were exposed to a longer separation (thirteen days). Almost all these infants were then studied for a further two years, namely until they were aged two and a half years, and their development was compared with that of a control sample of eight infants who had remained with mother throughout. A useful summary of findings from all these studies is given in Hinde & Spencer- Booth ( 1971).
Since findings on the initial four infants in regard to behaviour both during the separation and during the months after it are amply confirmed in the larger study, and since in the early publication by Spencer- Booth & Hinde ( 1967) behaviour is described in considerable detail, the account following is taken from that paper. Although, as in Kaufman's study, there was some variation in the reactions of the infants, they showed an overall pattern in common.
During the first day of separation all four infants screamed and geckered persistently. Although they did so less on subsequent days the amount continued much above what it had been before separation (and it continued so also for some weeks after mother's return). In contrast to Kaufman's pigtail infants, however, the rhesus infants were notably inactive immediately after mother's removal. This state of relative inactivity persisted throughout the remaining days of separation for all four infants:
-66-
'In general, the infants' behaviour during the mother's absence can only be described as depressed. They sat in the hunched, passive attitude of a subordinate animal. ' On the first day of separation both manipulative play and social play decreased dramatically. Although on subsequent days there was some recovery in manipulative play, social play remained low and tended to get progressively less during the six days.
Interaction of the infant with the adult male or one of the adult females of its group was greater during separation than either before or after it, but none the less was only a fraction of what the infant had been used to with its mother before separation. Here again there was much variation between infants. Whereas one separated infant was never observed being cradled by or clinging to an adult, each of the others was observed so for up to 20 per cent of the observation units. Nevertheless, such episodes were usually extremely brief in comparison with the long spells spent on mother before separation. Instead of a close cuddling the more normal picture was of a separated infant sitting touching or nearly touching aunts or the male; when adult moved away the infant would often whoo or gecker. Two of the infants, especially, often sat with the adult male of their group, running to him for protection. Thus the
62
separated infants received some substitute care from the other adults, but it was only a small proportion of what they had been receiving from their own mother.
The eating behaviour of the separated infants changed, too, in ways frequently seen in separated human children. On the first day of separation one infant ate hardly at all. Subsequently all four showed a tendency to eat more.
Although the acute disturbance of the first day subsided, then, the behaviour of all four infants remained far from normal throughout the subsequent five days of separation.
Disturbed behaviour was also much in evidence during the weeks after mother's return. After mother was returned to the. cage all four infants 'were more clinging than they had been before the separation. They showed exceptionally intense tantrums when rejected by their mothers, and often flung themselves violently on to their mothers, or sometimes, when the mother had rejected them, on to aunts. ' With two of the infants 'the effect was dramatic and long-lasting': they hardly came off their mothers during the first day of reunion.
A particularly striking feature during the first week or two of reunion was the way in which an infant might 'change from
-67-
being relaxed to being very upset and clinging without apparent cause'. On the second and third days after reunion one infant 'was recorded as coming off his mother in an apparently calm fashion, then suddenly panicking and going on her geckering'. Another infant (female) on the sixth day of reunion was observed to play in a relaxed way for half an hour and then to sleep on mother for a time: 'When she awoke she seemed very upset and terrified, cringed and would hardly leave her mother. ' Subsequently, however, she again seemed relaxed and began to play.
The behaviour of the four mother-infant pairs during the four months after reunion, up to the first birthday, varied much from pair to pair. Nevertheless, when the behaviour of the four infants was compared with that of a control group of eight infants who had had no separation experience, all four were found to have been affected, 'all taking the initiative in being close to their mothers more, relative to the controls, than before separation'. The tendency to cling and to stay close to mother was very marked; in one infant it persisted for the whole of the four months to the first birthday and in another for half that time.
Even more striking were the differences in behaviour shown by the four separated infants as compared with the eight controls when they were tested in slightly strange situations, first at twelve months of age and later at thirty months. The four who had been separated tended, in comparison with the controls, to be less willing to approach an experimeter offering food, to stay closer to mother when moved to a strange cage, to make shorter visits to a cage containing strange objects, and to be less active after having been frightened by a minor happening ( Hinde & Spencer-Booth 1968).
The results of these experiments, conducted almost two years after a separation lasting only six days, are impressive evidence that separation from mother can be traumatic. Whereas the behaviour of these young monkeys during the course of an uneventful day might show no differences from that of control monkeys, once the environment becomes a little unusual
63
differences show up: the behaviour of those who have been separated is more timid and anxious than is that of the controls. That this is true also of human infants has been noted on a number of occasions by Robertson ( 1953; 1958b) and the present writer ( Bowlby 1951; 1960a).
As remarked above, since their pioneer study of four rhesus
-68-
infants separated for six days Hinde & Spencer-Booth have done much further work and greatly amplified their findings. For those interested a description of some of the principal ones follows.
Further studies by Hinde and Spencer-Booth
From their extensive further studies Hinde & Spencer-Booth are able not only to confirm and amplify their data on the effects on young rhesus monkeys of a single six-day separation from mother but also to compare them with (a) the effects of a second short separation of six days and (b) the effects of a single, rather longer, separation of thirteen days. In addition, they are able to cast light on the factors responsible for the considerable degree of individual variation seen in the responses.
We consider, first, the short-term effects on the twice-separated infants; next, the long-term effects on infants who had been subjected to either one or two six-day separations; and, third, the short-term effects of a thirteen-day separation. (Long-term findings in respect of the last group of infants are not yet available. )
Short-term Effects of a Second Six-day Separation
When aged thirty to thirty-two weeks a total of eleven infants were separated from mother for six days for their first and only time and a total of ten for their second time. (Of the latter, five had had a first separation ten weeks earlier and five had had one five weeks earlier. ) When the behaviour of the infants separated for their second time was compared with that of infants of the same age separated for their first time no differences were apparent, on the measures used, either during the separation itself or during the subsequent month ( Spencer-Booth & Hinde 1971a). Evidence from the follow-up shows, however, that it would be a mistake to conclude that the effects of two separations are no different from the effects of one.
Long-term Effects of One and Two Six-day Separations
In their report on the long-term effects of six-day separations Spencer-Booth and Hinde ( 1971c) compare observations of infants from three samples: control infants, N = 8; infants that underwent a single six-day separation, N = 5; and infants that underwent two six-day separations, N = 8. 1 For some
____________________
1 Two infants first separated at twenty-one weeks of age, and one first separated at twenty-
six weeks, died before they were a year old, and a -69-
64
comparisons numbers are reduced; and on occasion it is necessary to pool data for the once- and the twice-separated infants. Stated briefly, findings are as follows:
1. When tested at twelve and again at thirty months of age and compared with the controls,
the previously separated infants of both groups still showed 'some persistence of those symptoms of depression and of disturbed mother-infant relations which had been conspicuous in the month immediately following the mother's return'.
2. The differences in behaviour between the previously separated infants and the controls were far more evident when the infant was tested in a strange environment than when it was in its home run.
3. Differences were much less marked at thirty months than at twelve months, but all the differences that were significant were in the direction that the previously separated animals showed poorer or more disturbed behaviour than the controls.
4. Most of the significant differences were between the controls and the twice-separated infants, with the once-separated infants generally occupying an intermediate position.
Let us consider these findings one by one.
In the home run at twelve months of age the previously separated animals of both groups tended to spend less time at a distance from mother, and to play a greater relative role in maintaining proximity to her, than did the control infants. Few of these differences were significant, however, and they were no longer observable either at eighteen or at thirty months. At all three ages, however, there was a significant tendency for the previously separated infants to show less locomotor activity and less social play than the controls.
Whereas the differences found in the behaviour of the previously separated animals and the controls while they were still in their home runs were not very conspicuous, when the animals were in a strange environment and tested there, differences were marked. The importance of this finding in its clinical implications can hardly be exaggerated.
____________________
second of the twenty-six-week group died soon after its second birthday. One of the long- separated group, together with three infants separated at eighteen weeks and omitted from consideration, also died before their first birthday. It is not clear how far the deprivation experience may have contributed to the death of these infants.
-70-
At the age of twelve months each infant with its mother was brought to a strange laboratory cage which communicated with a similar cage (the filter cage) by means of a passage large enough for the infant but not for the mother. Testing was conducted over a period of nine days and consisted of placing food or a strange object in the filter cage and seeing how the infant responded. Objects included a mirror, pieces of banana, and a yellow ball. On almost every test there was a significant tendency for the previously separated infants, in comparison with the controls, to wait longer before venturing to go alone into the filter cage, to pay shorter visits there, and to spend less total time in it. Furthermore, whenever scores of the twiceseparated infants differed from scores of the once-separated, divergence from the controls' scores was consistently greater in the case of the twice-separated infants. The table below, giving results for a test on the sixth day, when a yellow ball was placed in the filter cage, shows a typical pattern of differences.
65
Median score in minutes
Once- Twice- Controls separated separated
Measure N=6 N=5 N=8
Latency to entering
cage 0? 1 0? 1 0? 7 Total time in cage 7? 0 3? 9 3? 0 Median length of visit 0? 5 0? 3 0? 2 Time spent playing 2? 3 0 0
Another test given at twelve months
previously separated infants and the controls was one in which an infant was offered vitamins by an experimenter. The previously separated animals were much less willing to approach to obtain the vitamin than were the controls, and this was so even when the test was carried out in the home run. A probable explanation of that is that one of the experimenters had taken part in catching and removing the infants' mothers at the time of the separation.
Eighteen months later when aged thirty months each animal was given a series of comparable tests. On this occasion it was tested over a period of sixteen days when by itself in a screened
-71-
laboratory cage. Out of many tests only a few showed significant differences between groups. One was when the experimenter offered the vitamins. Another was when, on the second and sixth days, a piece of date was hung outside the cage just out of reach: the previously separated animals were much longer than the controls before making an attempt to get it, made fewer attempts, and persisted in making attempts for less long. (The number of animals available for this test made differentiation between once- and twice-separated animals impossible. )
Short-term Effects of a Thirteen-day Separation
At the age of thirty to thirty-two weeks a further six infants were separated for a single period lasting thirteen days (see Spencer-Booth & Hinde 1971b). During the whole of the second week they remained about as depressed and inactive as they were at the end of the first week. (This is in contrast to the moderate degree of recovery seen after the first week of separation in pigtail infants by Kaufman & Rosenblum. )
During the month following the separation it was found that the infants separated for thirteen days were significantly more affected than were infants in either of the other two separated groups. For at least the first week after reunion they emitted more distress calls; and
which showed significant differences between the
66
throughout the month they were more depressed. During the time they spent out of contact with mother they spent more time sitting about inactive than did the other separated infants, and when they were active they were less so. Whereas by the end of the first month of reunion infants separated once and for only six days were showing at least as much activity as they had done before the separation, the activity level of those separated for thirteen days was still significantly reduced. At this time the activity level of the twice-separated infants was intermediate between that of infants separated for a single six-day period and that of those separated for thirteen days.
From all these findings we can conclude with confidence not only that a single separation of no longer than six days at six months of age has perceptible effects two years later on rhesus infants, but that the effects of a separation are proportionate to its length. A thirteen-day separation is worse than a six-day; two six-day separations are worse than a single six-day separation. In these regards the effects of separation from mother can be
-72-
likened to the effects of smoking or of radiation. Although the effects of small doses appear negligible, they are cumulative. The safest dose is a zero dose.
Individual Variations of Response
There is much individual variation in the responses of rhesus infants to separation. Within the age-range studied age had little effect: whether a six-day separation was at twenty-one to twenty-two weeks, at twenty-five to twenty-six weeks, or at thirty to thirty-two weeks seemed to make little difference. Sex played some part: both during separation and after it males were more affected than females. Whether or not an infant was able to cling to another animal during separation had no effect on behaviour after reunion though such clinging did reduce the amount of distress calling at the time.
The most striking results to emerge from the analysis of the data on individual variation are the significant correlations between degree of distress shown by an infant and certain features of the mother-infant relationship ( Hinde & Spencer- Booth 1970). The infants that are most distressed during the first month after separation tend to be those that are most frequently rejected by mother and that play the greatest relative role in maintaining proximity to her. Since in regard to these features there is consistency for each mother-infant pair over time (as measured by rank-order correlations), it is not surprising to find that the degree of distress shown after separation is correlated both with the frequency with which a mother rejects her infant during the period before separation and with the frequency with which she rejects it during the period after reunion. In fact, it is found that, soon after mother returns, correlation of infant distress is higher with frequency of maternal rejection before separation than with frequency of rejection contemporaneously. Subsequently the balance changes and degree of distress becomes more highly correlated with the frequency with which she is rejecting her infant at the time.
Hinde & Spencer-Booth ( 1971) emphasize that these correlations do not justify the conclusion that differences in the mother-infant relation necessarily cause the differences in infants' responses to separations; nevertheless they believe that to be likely.
67
In a recent experiment Hinde & Davies ( 1972) altered the conditions in which separation occurs: instead of mothers being removed from the home cage and placed in a strange one,
-73-
the infants were removed and mothers remained behind. During their thirteen days of separation the behaviour of the five infants conformed to expectations. Although large individual differences made comparisons difficult, the infants separated in a strange cage seemed even more disturbed than did infants that remained in the home cage while mother was removed. After reunion with mother, however, the infants that had been separated in the strange cage were less disturbed than were infants whose mothers had been removed.
Certain observations of the behaviour of the mothers suggest an explanation of this unexpected finding. Compared with mothers that had been removed to the strange cage, mothers that remained behind were less distressed during their infant's absence and, after reunion, were more maternal and less rejecting of it; and harmonious interaction between the two was restored more quickly. These findings tend to support the view that a major determinant of the effect of a separation on a rhesus infant is how mother behaves towards it after reunion.
One form of behaviour that is extremely common in young children after a separation lasting a week or longer in strange surroundings and without substitute mothering, but that has only once been reported for monkey young, is detachment, namely a failure to recognize or respond to mother on reunion. In a study by Abrams (described by Mitchell 1970) twentyfour rhesus infants underwent a two-day separation from mother when aged between eight and twenty weeks. At the time of reunion one-quarter of the infants observed ran away from mother as she approached; and after a second two-day separation a few weeks later, the proportion that ran away doubled. Since, although on the lookout for detachment, Hinde & Spencer-Booth never observed it, the response may prove to be confined to infants in the very young age-range studied by Abrams. It is not yet clear, however, whether the response observed by Abrams can be regarded as homologous with that seen in young children.
The findings of the primate experiments have been described at length because they leave no serious doubt that most of what is to be seen during and after a brief separation in human infants is to be seen also in infants of other species. Explanations of human responses that presume cognitive processes at a specifically human level are thus called in question.
-74-
68
Part II
AN ETHOLOGICAL APPROACH TO HUMAN FEAR
-75-
Chapter 5
Basic Postulates in Theories of Anxiety and Fear
Paradigms provide scientists not only with a map but also with some of the directions essential for map-making. In learning a paradigm the scientist acquires theory, methods and standards together, usually in an inextricable mixture. . . . That is [a] reason why schools guided by different paradigms are always slightly at crowpurposes.
THOMAS S. KUHN ( 1962)
Anxiety allied to fear
Though at intervals down the years one student of the problem after another has been struck that a principal source of anxiety and distress is separation from loved figures, or the threat of separation, there has been great reluctance to accept that simple formula. Objections to it are extremely deep-seated and based on one or more of several common assumptions each of which, it is argued here, is tenable no longer.
In this and the following chapters the simple view is once again advanced. And because there has been so much incredulity and therefore reasoned opposition to it, it is presented in some detail. We begin by considering some of the common assumptions that underlie the traditional incredulity and opposition, with special reference to the influence of Freud's early formulated theory of motivation.
In all psychoanalytic and psychiatric discussion of anxiety it is taken for granted that the emotional states referred to respectively as 'anxiety' and 'fear' are closely related. In just what way they are related is the puzzle. Freud is concerned repeatedly both to compare and to contrast the two: see, for example, Addendum B to Inhibitions, Symptoms and Anxiety ( SE 20: 164168). Others have followed in his steps. Reviewing the whole confused scene in a recent article, Lewis ( 1967) emphasizes that, throughout the broad field of psychopathology, the word 'anxiety' is used habitually to refer to 'an emotional state with
-77-
the subjectively experienced quality of fear or a closely related emotion'. Often, we know, the two words are used interchangeably. In view of the close relatedness of the emotional states concerned and also of the meanings of the two words, it is hardly surprising that ideas about the conditions that give rise to the one state should influence ideas about the conditions that give rise to the other.
69
Nevertheless, in all this diverse, confusing, and contradictory theorizing there is one matter on which all seem agreed: whereas the nature and origin of anxiety are obscure, the nature and origin of fear are simple and readily intelligible.
In the theories advanced in this work there is only one break with that tradition. As hitherto, the states referred to by the words 'anxiety' and 'fear' are seen as closely related. Furthermore, ideas about what arouse states of the one kind are, also as hitherto, closely linked to ideas about what arouse states of the other. Where the paths diverge is at an altogether different point, namely in theories about the nature of the conditions that are apt to arouse what is allegedly the more easily understood of the two kinds of state, namely that of fear itself.
In psychoanalytic and psychiatric circles, it is argued, there still flourish seriously misconceived assumptions about fear and the conditions that arouse it. These mistaken assumptions have long had, and continue to have, a most adverse effect on our ability to understand the distressing anxieties and fears from which our patients suffer.
A further finding from these miniature separation experiments, and one that links with the findings of Shirley ( 1942) and Heathers ( 1954) (see pp. 35 and 37 above), is only very recently reported. This is that when a child of about one year is tested in Ainsworth's series of episodes for a second time, a few weeks after the first testing, he is more upset and anxious than he was on the first occasion. When mother is present he keeps closer to her and clings more tightly. When she is absent he cries more ( Ainsworth, personal communication). These findings emerge from a test--retest study of twenty-four babies tested first at fifty weeks of age and a second time two weeks later. On the assumption that increased sensitivity is not due simply to maturation, which is unlikely, these findings provide the first experimental evidence that at one year of age a separation lasting only a few minutes, in what would ordinarily be regarded as a bland situation, is apt to leave a child more sensitive than he was before to a repetition of the experience.
Ontogeny of responses to separation The First Year
Since the responses to separation that are so unmistakable in infants of twelve months and older are not present at birth, it is clear that they must develop at some time during the first year of life. Unfortunately, studies designed to throw light on this development are few, and are confined to infants admitted to hospital. Nevertheless such evidence as is available is unambiguous. It is in keeping, moreover, with what is known about the
-52-
development of attachment behaviour and about cognitive development generally.
In Chapter 15 of the earlier volume the steps by which, during the early months of life, an infant's attachment behaviour gradually becomes focused on a discriminated and preferred attachment figure are described. Development can be summarized as follows: before sixteen weeks differentially directed responses are few in number and are seen only when methods of observation are sensitive; between sixteen and twenty-six weeks differentially directed responses are both more numerous and more apparent; and in the great majority of family infants of six months and over they are plain for all to see. It comes as no surprise, therefore, that the full range of responses to separation described in earlier sections of this chapter is not seen before six or seven months of age.
Schaffer studied seventy-six infants of various ages under twelve months admitted to hospital: none was marasmic, deformed, or thought to be brain-damaged. Of the total, twentyfive were
51
healthy infants admitted for elective surgery. While in hospital each child was observed during a two-hour session on each of the first three days (see Schaffer 1958; Schaffer & Callender 1959). Infants were not only without mother but had very little social interaction with nurses.
The responses observed in these twenty-five healthy infants differed greatly according to the child's age. The dividingpoint was twenty-eight weeks. Of the sixteen aged twenty-nine weeks and over, all but one fretted piteously, exhibiting all the struggling, restlessness, and crying so typical of two- and threeyear-olds. Of the nine aged twenty-eight weeks and under, by contrast, all but two 1 are reported to have accepted the situation without protest or fretting: only an unwonted and bewildered silence indicated their awareness of change.
Schaffer emphasizes that the shift from a bewildered response to active protest and fretting occurs suddenly and at full intensity at about twenty-eight weeks af age. Thus, of the sixteen infants aged between twenty-nine and fifty-one weeks, both the length of the period of fretting and the intensity of it were as great in those of seven and eight months as in those of eleven and twelve months.
Furthermore, responses both to the observers and to mother when she visited changed equally suddenly at about thirty weeks:
____________________
1 One of the exceptions was an infant already twenty-eight weeks of age.
-53-
In the younger infants [twenty-eight weeks and less] most observation sessions showed these infants to be normally responsive, although the people confronting them were complete strangers. This held as much for the nurses who fed and bathed them as for the observers. . . . In the older group, on the other hand, normal [i. e. friendly] responsiveness was almost completely lacking, and the majority of observation sessions showed these infants to be negative and frightened when approached by a stranger-- a type of behaviour not seen at all in the younger group ( Schaffer & Callender 1959).
Although there were too few observations for statistical comparisons to be made of how infants of different ages responded to mother's visit, such observations as could be made support the thesis of a sharp change around twenty-eight to thirty weeks. Infants older than that mostly clung rather desperately to mother, behaviour that was in striking contrast to their negative responses to the observers. The younger infants, by contrast, tended to respond to mother and to observers without showing marked discrimination between them. Similarly, when mother departed, whereas older infants cried loudly and for a long time, even desperately, the younger ones showed no sign of protest.
Finally, the behaviour of the infants on their return home from hospital differed greatly according to age-group. Most of the infants aged seven months and over showed intense attachment behaviour. They clung almost continuously to mother, cried loud and long when left alone by her, and were notably afraid of strangers. Even figures formerly familiar, such as father and siblings, were sometimes regarded with suspicion. Infants aged under seven months, by contrast, showed little or no attachment behaviour during their early days at home. Their mothers described them as 'strange'. On the one hand, these infants seemed utterly
52
preoccupied in scanning the environment; on the other, they seemed unheedful of adults or perhaps averted their head when approached:
For hours on end sometimes the infant would crane his neck, scanning his surroundings without apparently focusing on any particular feature and letting his eyes sweep over all objects without attending to any particular one. A completely blank expression was usually observed on his face, though sometimes a bewildered or frightened look was reported. In the extreme form of this syndrome the infants were quite inactive throughout, apart from the scanning behaviour, and
-54-
no vocalization was heard though one or two were reported to have cried or whimpered. When confronted with a toy the infant disregarded it.
To attempts by adults to make contact with them some of these younger infants seemed altogether oblivious. Others seemed to avoid the adults, and others again to gaze 'through' them with the same blank look that they used for the rest of the environment.
The only way in which the responses of the infants of the two age-groups were similar was in regard to sleep: in infants of both groups disturbed sleep and night-crying were common.
How the responses of infants of under seven months are best understood, and what their significance for an infant's future development may be, is difficult to know. It is plain, however, that the responses of these younger infants to separation are very different at every phase from those of older ones, and that it is only after about seven months of age that the patterns that are the subject of this work are to be seen.
In discussion of his findings Schaffer ( 1958) draws on Piaget's work on the development of an infant's concept of an object ( Piaget 1937). Only during the second half of the first year, Piaget finds, is there evidence that an infant is beginning to be able to conceive of an object as something that exists independently of himself, in a context of spatial and causal relations, even when it is not present to his perception, and so to search for it when it is missing. Bell ( 1970) confirms Piaget's findings and, in addition, reports the results of an experiment designed to test whether or not an infant develops a capacity to conceive of a person as a persisting object earlier than he develops the capacity in regard to inanimate things. Although her results show that a majority of infants develop the capacity in regard to a person earlier than in regard to things, it is not until about the ninth month that the capacity in regard to persons is reasonably well developed, and in a minority it lags some weeks behind that. For reasons to do with cognitive development, therefore, the types of response to separation with which we are concerned could hardly be expected in infants younger than those in whom they are seen.
Change after the First Birthday
All the evidence suggests that, once established, the typical patterns of response to being placed in strange surroundings
-55-
53
with strange people do not undergo marked change, either in form or in intensity, much before the third birthday. Thenceforward their intensity begins to diminish, but only slowly. As an example, the change in Lottie's feelings about going to nursery school that appeared at the beginning of the third term, when she was aged two years and nine months (see above, p. 38 ), is characteristic of many children. Provided he knows where his mother is and has good reason to expect her to return soon, a child begins to accept another fairly familiar person, even when he is in a fairly unfamiliar place.
The only conditions at present known that reduce appreciably the effects of separation from mother are familiar possessions, the companionship of another and familiar child and, as Robertson & Robertson ( 1971) have shown, especially mothering from a skilled and familiar foster mother. By contrast, strange people, strange places, and strange proceedings are always alarming; and they are especially alarming when encountered alone (see Chapters 7 and 8). ?
Since distress at being separated unwillingly from an attachment figure is an indissoluble part of being attached to someone, changes occurring with age in the form of response to separation accompany, step by step, changes in the form that attachment behaviour takes. These changes are sketched in the first volume (Chapters 11 and 17) and need not be described further here. In so far as attachments to loved figures are an integral part of our lives, a potential to feel distress on separation from them and anxiety at the prospect of separation is so also. That is a theme running through the rest of this volume.
Meanwhile, in order that we may view the responses to separation seen in humans in a perspective broader than has been traditional, it is useful to compare the responses of young human children with those of the young of other species. When that is done it becomes evident that, just as attachment behaviour occurs in rather similar forms across a number of mammalian and avian species, so also do responses to separation. Here again man is no isolated case.
-56-
54
Chapter 4
Behaviour with and without Mother: Non-human Primates
Man with all his noble qualities, with sympathy that feels for the most debased, with benevolence which extends not only to other men but to the humblest living creature, with his god-like intellect which has penetrated into the movements and constitution of the solar system -- with all these exalted powers -- still bears in his bodily frame the indelible stamp of his lowly origin.
CHARLES DARWIN ( 1871)
Naturalistic observations
It has long been known that isolation and separation from a mother figure can cause distress, expressed in calling and searching, in the young of many species of bird and mammal. The 'lost piping' of young ducklings who have become attached to and have temporarily lost a mother figure is a familiar example. Others are the bleating of lambs and the yelping of puppies. Coming nearer man, there are numerous examples in the accounts of monkey and ape infants brought up by human caretakers. All accounts agree on the intensity of protest exhibited whenever a baby primate loses its mother figure, and the intensity of distress that follows when she cannot be found. All agree, too, on the intensity of clinging that occurs after the two are reunited.
For example, Bolwig ( 1963), in his account of the little patas monkey he reared from a few days old,1 describes how from the first the little monkey 'showed no fear of man, cried much, and panicked when left alone. . . . The screaming, with wide open mouth and distorted face, was only heard when the observer moved out of the monkey's immediate reach or sight. On such occasions he would more often than not run staggering to the nearest person in sight. ' Soon the monkey's attachment had become focused on Bolwig himself and then, until three and a
____________________
1 See Volume I, Chapter 11.
-57-
half months of age, the monkey could be very troublesome unless perpetually with his caretaker.
By the age of four months, however, the little monkey was exploring increasingly far afield and his master decided
to leave him for some hours every day in a cage with other monkeys of his own kind. This attempt was, however, not very successful. Although he knew the other monkeys well and was accustomed to play with them he panicked as soon as he knew I wanted to leave him
55
behind, screamed, clung desperately to me and then tried to tear the door open. He would sit and cry until I finally let him out. Afterwards he would cling to me and refuse to leave me out of sight for the rest of the day. In the evening when asleep he would wake up with small shrieks and cling to me, showing all signs of terror when I tried to release his grip.
Accounts of similar behaviour are given of infant chimpanzees. Cathy Hayes ( 1951) recounts how Viki, a female she adopted at three days, would, when aged four months, cling to her foster mother
from the moment she left her crib until she was tucked in at night. . . . She sat on my lap while I ate or studied. She straddled my hip as I cooked. If she were on the floor, and I started to get away, she screamed and clung to my leg until I picked her up. . . . If some rare lack of vigilance on her part let a room's length separate us, she came charging across the abyss, screaming at the height of her considerable ability.
The Kelloggs, who did not adopt their female chimp, Gua, until she was seven months old and who kept her for nine months, report identical behaviour ( Kellogg & Kellogg 1933). They describe
an intensive and tenacious impulse to remain within sight and call of some friend, guardian, or protector. Throughout the entire nine months . . . whether indoors or out, she almost never roamed very far from someone she knew. To shut her up in a room by herself, or to walk away faster than she could run, and to leave her behind, proved, as well as we could judge, to be the most awful punishment that could possibly be inflicted. She could not be alone apparently without suffering.
Comparing Gua with their son, who was two and a half months older than she, the Kelloggs report:
-58-
Both subjects displayed what might be called anxious behaviour (i. e. fretting and crying) if obvious preparations were being made by the grown-ups to leave the house. This led (in Gua) to an early understanding of the mechanism of door closing and a keen and continual observation of the doors in her vicinity. If she happened to be on one side of a doorway, and her friends on the other, the slightest movement of the door toward closing, whether produced by human hands or by the wind, would bring Gua rushing through the narrowing aperture, crying as she came.
The very detailed observations made by van Lawick-Goodall ( 1968) of chimpanzees in the Gombe Stream Reserve in central Africa show not only that anxious and distressed behaviour on being separated, as reported of animals in captivity, occurs also in the wild but that distress at separation continues throughout chimpanzee childhood. During the first year an infant is rarely out of actual contact with mother and, although from its first birthday onwards it spends more time out of contact, it none the less remains in proximity to her. Not until young are four and a half years of age are any of them seen travelling not in the company of mother, and then only rarely. 1
56
Once an infant begins to spend time out of contact with mother, proximity is maintained largely by auditory signals. Either mother or infant emits a 'hoo' whimper to which the other promptly responds:
When the infant . . . begins to move from its mother, it invariably utters this sound if it gets into any difficulty and cannot quickly return to her. Until the infant's locomotor patterns are fairly well developed the mother normally responds by going to fetch it at once. The same sound is used by the mother when she reaches to remove her infant from some potentially dangerous situation or even, on occasion, as she gestures it to cling on when she is ready to go. The 'hoo' whimper therefore serves as a fairly specific signal in re-establishing mother-- infant contact.
Another signal used by infants is a scream; it is elicited whenever an infant falls or nearly falls from its mother or is frightened by a sudden loud noise. When her infant screams a mother almost unfailingly retrieves it and cradles it: 'On several occasions
____________________
1 For a brief description of the developmental course of attachment behaviour in
chimpanzees see Volume I. Chanter 11. -59-
infants screamed loudly when their mothers started to move away without them. Each time the mothers immediately turned back and retrieved them. Indeed, throughout infancy, screaming normally results in the mother hurrying to rescue her child. '
Juveniles up to five or six years old also scream when lost or in trouble, and again mother usually hurries to the rescue:
On several occasions juveniles were observed who had accidentally lost their mothers. In each instance, after peering round from various trees, whimpering and screaming as they did so, they hurried off -- often in the wrong direction. On three occasions I was able to observe the reaction of the mother and every time, although she set off in the direction of her offspring's screaming, she herself made no sound to indicate her whereabouts.
In one case a juvenile female aged five years lost her mother in the evening and was still whimpering and crying the following morning. In another case, a juvenile stopped screaming before her mother found her, which resulted in a separation lasting several hours. (No information is available regarding behaviour of the young after reunion. )
Thus in these wild-living chimpanzees proximity of young to mother is maintained until pre- adolescence. Separations are rare, and usually quickly rectified by vocal signals and mutual search.
Early experimental studies
These naturalistic accounts show plainly not only that the attachment behaviour of young non- human primates is very similar to the attachment behaviour of young children but that their responses to separation are very similar also. Because of this, and because experimental separations lasting longer than minutes are inadmissible in the case of human young, more
57
than one scientist has turned to monkey young for experimental subjects. A number of studies from at least four different centres are now published. Animals used include infants aged between two and eight months, of five different species, namely four species of macaque (the rhesus, the pigtail, the bonnet, and the Java) and the patas monkey. All five are species of semiterrestrial and group-living old-world monkeys. 1
____________________
1 The nearest experiments so far reported with chimpanzees are those by Mason ( 1965), but
in this case the separations were from a cage-mate of the same age and not from mother. -60-
Responses to separation differ as between species, though differences are more of intensity than of kind. In the case of rhesus, pigtail, and Java macaques great distress is observed throughout the period of separation itself and, afterwards, there is a very marked tendency to cling to mother and to resist any attempt at a further separation, however brief. In the case of both bonnet macaques and patas monkeys, intense distress is again seen during the first hours after separation, but then it wanes; thereafter activity is less depressed than in the other species of macaque and there is much less disturbance after reunion with mother. The reduction of distress in the bonnet macaques appears to come about in great part because the separated infant receives continuous substitute care from one of the other familiar females in the group.
In what follows attention is given to the studies using rhesus and pigtail infants both because their responses appear to resemble more closely those of human infants and because the studies of these species are more numerous and extensive, especially in the case of the rhesus. Those wishing to compare the behaviour of bonnet macaques are referred to the study by Rosenblum & Kaufman ( 1968; see also Kaufman & Rosenblum 1969); and of patas monkeys to the study by Preston, Baker & Seay ( 1970). A useful review of separation studies is given by Mitchell ( 1970).
An early experimental study was carried out by Jensen & Tolman ( 1962). When two infant pigtail monkeys, each reared in a cage alone with mother, were aged respectively five and seven months, the infants and mothers were exchanged on several occasions for periods of no longer than five minutes. Observation was through a one-way vision screen.
Because mother and infant cling tightly to one another separation cannot be achieved with monkeys except by deception or by the exercise of a good deal of force. The protests of both parties are intense. Jensen & Tolman give a vivid account:
Separation of mother and infant monkeys is an extremely stressful event for both mother and infant as well as for the attendants and for all other monkeys within sight or earshot of the experience. The mother becomes ferocious toward attendants and extremely protective of her infant. The infant's screams can be heard over almost the entire building. The mother struggles and attacks the separators. The baby clings tightly to the mother and to any object which it can
-61-
58
grasp to avoid being held or removed by the attendant. With the baby gone, the mother paces the cage almost constantly, charges the cage occasionally, bites at it, and makes continual attempts to escape. She also lets out occasional mooing-like sounds. The infant emits high pitched shrill screams intermittently and almost continuously for the period of separation.
As soon as the five minutes had elapsed and mother and infant were reunited each immediately went to the other and the two remained in the closest possible contact: 'The mother sits quietly holding her baby, and if no attendants are present she very quickly seems content and relaxed. All is quiet in the room. No more piercing screeches of the baby or sounds from the mother are heard. ' The duration of this unbroken period of intense mutual clinging following a separation that had lasted a mere five minutes was never less than fifteen minutes and in some cases as long as forty.
Other workers have subjected their monkey infants to much longer separations, the periods ranging from six days to as long as four weeks. In the case of pigtail and rhesus infants all observers report extreme and noisy distress during the twentyfour hours or so immediately after separation followed by a quieter period of a week or more during which the infants show little activity or play and, instead, sit hunched up and depressed.
Harlow has been responsible for two such studies. In one ( Seay, Hansen & Harlow 1962), four rhesus infants, ages ranging from twenty-four to thirty weeks, were kept apart from mother for a period of three weeks. 1 Since mother was in an adjacent cage and only a transparent screen separated the two, each could see and hear the other. Observations were made at regular intervals during the three weeks prior to separation, during the three weeks of separation, and for three weeks following separation. On each occasion two infants, already familiar with each other, were separated simultaneously and, during the period of separation, each infant had free access to the other. Thus throughout the period of separation all four infants had companionship, access to food and water,
____________________
1 For an account of attachment behaviour in rhesus monkeys see Volume I, Chapter 11.
Until it is about three years of age a young rhesus monkey in the wild remains close to mother.
-62-
and also visual and auditory contact with mother. Only physical contact with mother was missing.
As soon as the transparent screen had been lowered all four infants engaged in 'violent and prolonged protest'. There was much high-pitched screeching and crying, they made numerous attempts to reach mother, including hurling themselves against the screen, and they also scampered in a disoriented way around the cage. Later, when quiet, the infants huddled against the screen in as close proximity to mother as they could get. Initially mothers barked and threatened the experimenter, but their responses were less intense and persistent than the infants'. Throughout the separation period the pairs of separated infants showed little interest in one another and little play, in contrast to the active play between them seen in the three weeks prior to separation and after it was over. In the days after mother and infant had been reunited there was a very marked increase in the incidence of infant clinging to mother and
59
keeping in contact with her compared with what had occurred during the days before separation.
In a second and similar experiment Seay & Harlow ( 1965) separated a further eight rhesus infants from mother when the infants were aged thirty weeks. On this occasion the separation lasted only two weeks and mother was removed completely; a second separated infant was again available for play, but this time for only half an hour each day. Results were as before, including on the first day 'disoriented running about, climbing, screeching and crying'; relatively little interest was shown in the companion infant. After the phase of protest (duration of which is not reported) the infants 'passed into a stage characterized by low activity, little or no play and occasional crying'. The writers express the belief that 'this second stage is behaviorally similar to that described as despair of children separated from their mothers'. Immediately after reunion once again there was a phase of strong mutual clinging by mother and infant.
In 1966 and 1967 reports were published of closely similar findings by two other groups of workers, Spencer- Booth & Hinde using rhesus macaques in Cambridge, Kaufman & Rosenblum using pigtail macaques in New York. These two studies have much in common and are more informative than the earlier ones.
Whereas in both Jensen's and Harlow's laboratories the infants had been brought up with mother alone, each pair in a small cage, in the laboratories of Hinde
-63-
and of Kaufman infant and mother lived as part of a stable social group in a fairly large cage. Present with them were an adult male, two or three other adult females, and often other young. In both laboratories separation was effected by removing mother from the cage. This meant that the infant remained behind in a completely familiar environment with a number of other familiar animals: the only change in its life was the absence of mother.
A second advantage of the Hinde and the Kaufman studies is that results are reported in much greater detail, both in regard to the course of behaviour during the week or more of separation and in regard to the behaviour of both partners during the period of months, and in the Hinde study of nearly two years, after the separation was over. These observations are especially valuable in giving information about subsequent effects of the experimental separation.
In the experiment of Kaufman & Rosenblum ( 1967) the subjects were four infant pigtail monkeys ranging in age from twenty-one to twenty-six weeks. In each case mother was removed from the cage for a period of four weeks. The behaviour seen during the separation is reported to have fallen into three phases, described as 'agitation, depression and recovery'. Whereas three infants went through all three phases, the fourth, the daughter of the dominant female of the group, showed comparatively little depression and spent much of her time with the other adult females of her group. The behaviour of the remaining three infants is described as follows:
During the first phase pacing, searching head movements, frequent trips to the door and windows, sporadic and shortlived bursts of erratic play, and brief movements toward other members of the group seemed constant. Cooing, the rather plaintive distress call of the young macaque, was frequent. There was an increased amount of self-directed behaviour, such as sucking of digits, and mouthing and handling of other parts of the body, including the
60
genitals. The reaction persisted throughout the first day, during which time the infant did not sleep.
After 24 to 36 hours the pattern in three infants changed strikingly. Each infant sat hunched over, almost rolled into a ball, with his head often down between his legs. Movement was rare except when the infant was actively displaced. The movement that did occur appeared to be in slow motion,
-64-
except at feeding time or in response to aggression. The infant rarely responded to social invitation or made a social gesture, and play behaviour virtually ceased. The infant appeared disinterested in and disengaged from the environment. Occasionally he would look up and coo.
After persisting unchanged for 5 or 6 days the depression gradually began to lift. The recovery started with a resumption of a more upright posture and a resurgence of interest in the inanimate environment. Slow tentative exploration appeared with increasing frequency. Gradually, the motherless infant also began to interact with his social environment primarily with his peers, and then he began to play once again. The depression continued, but in an abated form. Periods of depression alternated with periods of inanimate-object exploration and play. Movement increased in amount and tempo. Toward the end of the month the infant appeared alert and active a great deal of the time, yet he still did not behave like a typical infant of that age.
Throughout the three months following mother's return the behaviour of mother and infant was recorded. In all four cases significant changes were observed, similar in direction to those observed in earlier studies:
When the mother was reintroduced to the group another dramatic change occurred. There was a tremendous reassertion of the dyadic relationship with marked increases in various measures of closeness in all four pairs. Clinging by the infant, protective enclosure by the mother, and nipple contact all rose significantly in the month after the reunion as compared to the frequency of these actions in the month before separation. Even in the third month after the reunion this trend was evident. This significant rise in measures of dyadic closeness is particularly striking in view of the fact that ordinarily for the age periods involved these particular behaviours fall considerably.
The increased closeness was manifest in other ways as well. A measure of mother--infant physical separation that we have found valuable in our normative studies concerns departures (usually by the infant) to another level of the pen. The frequency of such departures during the month after the reunion fell to 20 per cent of the departures in the month before the separation. Furthermore, the mean duration of these departures fell from 60? 5 seconds to 34? 4 seconds.
-65-
Not only did infants depart less frequently and for shorter periods than before the separation had occurred but mothers were more tolerant than before of their infant's persistent proximity, only rarely discouraging it by rejection or withdrawal.
61
Compared with the four subjects of Kaufman's experiment, the four subjects described initially by Spencer-Booth and Hinde were not only of a different species (rhesus instead of pigtail infants) but were rather older (thirty to thirty-two weeks of age instead of twenty-one to twenty-six weeks); and the length of time for which mother was removed was much shorter -- only six days in place of four weeks. Nevertheless, the behaviour observed both during the separation and during the months afterwards closely resembled what Kaufman and his colleagues were seeing (see Hinde, Spencer- Booth & Bruce 1966; Spencer- Booth & Hinde 1967).
Since their first reports Spencer-Booth and Hinde have published findings from a series of further studies in which the initial sample of four infants was increased to twenty-one, infants in certain sub-samples were exposed to a second separation, and another six infants were exposed to a longer separation (thirteen days). Almost all these infants were then studied for a further two years, namely until they were aged two and a half years, and their development was compared with that of a control sample of eight infants who had remained with mother throughout. A useful summary of findings from all these studies is given in Hinde & Spencer- Booth ( 1971).
Since findings on the initial four infants in regard to behaviour both during the separation and during the months after it are amply confirmed in the larger study, and since in the early publication by Spencer- Booth & Hinde ( 1967) behaviour is described in considerable detail, the account following is taken from that paper. Although, as in Kaufman's study, there was some variation in the reactions of the infants, they showed an overall pattern in common.
During the first day of separation all four infants screamed and geckered persistently. Although they did so less on subsequent days the amount continued much above what it had been before separation (and it continued so also for some weeks after mother's return). In contrast to Kaufman's pigtail infants, however, the rhesus infants were notably inactive immediately after mother's removal. This state of relative inactivity persisted throughout the remaining days of separation for all four infants:
-66-
'In general, the infants' behaviour during the mother's absence can only be described as depressed. They sat in the hunched, passive attitude of a subordinate animal. ' On the first day of separation both manipulative play and social play decreased dramatically. Although on subsequent days there was some recovery in manipulative play, social play remained low and tended to get progressively less during the six days.
Interaction of the infant with the adult male or one of the adult females of its group was greater during separation than either before or after it, but none the less was only a fraction of what the infant had been used to with its mother before separation. Here again there was much variation between infants. Whereas one separated infant was never observed being cradled by or clinging to an adult, each of the others was observed so for up to 20 per cent of the observation units. Nevertheless, such episodes were usually extremely brief in comparison with the long spells spent on mother before separation. Instead of a close cuddling the more normal picture was of a separated infant sitting touching or nearly touching aunts or the male; when adult moved away the infant would often whoo or gecker. Two of the infants, especially, often sat with the adult male of their group, running to him for protection. Thus the
62
separated infants received some substitute care from the other adults, but it was only a small proportion of what they had been receiving from their own mother.
The eating behaviour of the separated infants changed, too, in ways frequently seen in separated human children. On the first day of separation one infant ate hardly at all. Subsequently all four showed a tendency to eat more.
Although the acute disturbance of the first day subsided, then, the behaviour of all four infants remained far from normal throughout the subsequent five days of separation.
Disturbed behaviour was also much in evidence during the weeks after mother's return. After mother was returned to the. cage all four infants 'were more clinging than they had been before the separation. They showed exceptionally intense tantrums when rejected by their mothers, and often flung themselves violently on to their mothers, or sometimes, when the mother had rejected them, on to aunts. ' With two of the infants 'the effect was dramatic and long-lasting': they hardly came off their mothers during the first day of reunion.
A particularly striking feature during the first week or two of reunion was the way in which an infant might 'change from
-67-
being relaxed to being very upset and clinging without apparent cause'. On the second and third days after reunion one infant 'was recorded as coming off his mother in an apparently calm fashion, then suddenly panicking and going on her geckering'. Another infant (female) on the sixth day of reunion was observed to play in a relaxed way for half an hour and then to sleep on mother for a time: 'When she awoke she seemed very upset and terrified, cringed and would hardly leave her mother. ' Subsequently, however, she again seemed relaxed and began to play.
The behaviour of the four mother-infant pairs during the four months after reunion, up to the first birthday, varied much from pair to pair. Nevertheless, when the behaviour of the four infants was compared with that of a control group of eight infants who had had no separation experience, all four were found to have been affected, 'all taking the initiative in being close to their mothers more, relative to the controls, than before separation'. The tendency to cling and to stay close to mother was very marked; in one infant it persisted for the whole of the four months to the first birthday and in another for half that time.
Even more striking were the differences in behaviour shown by the four separated infants as compared with the eight controls when they were tested in slightly strange situations, first at twelve months of age and later at thirty months. The four who had been separated tended, in comparison with the controls, to be less willing to approach an experimeter offering food, to stay closer to mother when moved to a strange cage, to make shorter visits to a cage containing strange objects, and to be less active after having been frightened by a minor happening ( Hinde & Spencer-Booth 1968).
The results of these experiments, conducted almost two years after a separation lasting only six days, are impressive evidence that separation from mother can be traumatic. Whereas the behaviour of these young monkeys during the course of an uneventful day might show no differences from that of control monkeys, once the environment becomes a little unusual
63
differences show up: the behaviour of those who have been separated is more timid and anxious than is that of the controls. That this is true also of human infants has been noted on a number of occasions by Robertson ( 1953; 1958b) and the present writer ( Bowlby 1951; 1960a).
As remarked above, since their pioneer study of four rhesus
-68-
infants separated for six days Hinde & Spencer-Booth have done much further work and greatly amplified their findings. For those interested a description of some of the principal ones follows.
Further studies by Hinde and Spencer-Booth
From their extensive further studies Hinde & Spencer-Booth are able not only to confirm and amplify their data on the effects on young rhesus monkeys of a single six-day separation from mother but also to compare them with (a) the effects of a second short separation of six days and (b) the effects of a single, rather longer, separation of thirteen days. In addition, they are able to cast light on the factors responsible for the considerable degree of individual variation seen in the responses.
We consider, first, the short-term effects on the twice-separated infants; next, the long-term effects on infants who had been subjected to either one or two six-day separations; and, third, the short-term effects of a thirteen-day separation. (Long-term findings in respect of the last group of infants are not yet available. )
Short-term Effects of a Second Six-day Separation
When aged thirty to thirty-two weeks a total of eleven infants were separated from mother for six days for their first and only time and a total of ten for their second time. (Of the latter, five had had a first separation ten weeks earlier and five had had one five weeks earlier. ) When the behaviour of the infants separated for their second time was compared with that of infants of the same age separated for their first time no differences were apparent, on the measures used, either during the separation itself or during the subsequent month ( Spencer-Booth & Hinde 1971a). Evidence from the follow-up shows, however, that it would be a mistake to conclude that the effects of two separations are no different from the effects of one.
Long-term Effects of One and Two Six-day Separations
In their report on the long-term effects of six-day separations Spencer-Booth and Hinde ( 1971c) compare observations of infants from three samples: control infants, N = 8; infants that underwent a single six-day separation, N = 5; and infants that underwent two six-day separations, N = 8. 1 For some
____________________
1 Two infants first separated at twenty-one weeks of age, and one first separated at twenty-
six weeks, died before they were a year old, and a -69-
64
comparisons numbers are reduced; and on occasion it is necessary to pool data for the once- and the twice-separated infants. Stated briefly, findings are as follows:
1. When tested at twelve and again at thirty months of age and compared with the controls,
the previously separated infants of both groups still showed 'some persistence of those symptoms of depression and of disturbed mother-infant relations which had been conspicuous in the month immediately following the mother's return'.
2. The differences in behaviour between the previously separated infants and the controls were far more evident when the infant was tested in a strange environment than when it was in its home run.
3. Differences were much less marked at thirty months than at twelve months, but all the differences that were significant were in the direction that the previously separated animals showed poorer or more disturbed behaviour than the controls.
4. Most of the significant differences were between the controls and the twice-separated infants, with the once-separated infants generally occupying an intermediate position.
Let us consider these findings one by one.
In the home run at twelve months of age the previously separated animals of both groups tended to spend less time at a distance from mother, and to play a greater relative role in maintaining proximity to her, than did the control infants. Few of these differences were significant, however, and they were no longer observable either at eighteen or at thirty months. At all three ages, however, there was a significant tendency for the previously separated infants to show less locomotor activity and less social play than the controls.
Whereas the differences found in the behaviour of the previously separated animals and the controls while they were still in their home runs were not very conspicuous, when the animals were in a strange environment and tested there, differences were marked. The importance of this finding in its clinical implications can hardly be exaggerated.
____________________
second of the twenty-six-week group died soon after its second birthday. One of the long- separated group, together with three infants separated at eighteen weeks and omitted from consideration, also died before their first birthday. It is not clear how far the deprivation experience may have contributed to the death of these infants.
-70-
At the age of twelve months each infant with its mother was brought to a strange laboratory cage which communicated with a similar cage (the filter cage) by means of a passage large enough for the infant but not for the mother. Testing was conducted over a period of nine days and consisted of placing food or a strange object in the filter cage and seeing how the infant responded. Objects included a mirror, pieces of banana, and a yellow ball. On almost every test there was a significant tendency for the previously separated infants, in comparison with the controls, to wait longer before venturing to go alone into the filter cage, to pay shorter visits there, and to spend less total time in it. Furthermore, whenever scores of the twiceseparated infants differed from scores of the once-separated, divergence from the controls' scores was consistently greater in the case of the twice-separated infants. The table below, giving results for a test on the sixth day, when a yellow ball was placed in the filter cage, shows a typical pattern of differences.
65
Median score in minutes
Once- Twice- Controls separated separated
Measure N=6 N=5 N=8
Latency to entering
cage 0? 1 0? 1 0? 7 Total time in cage 7? 0 3? 9 3? 0 Median length of visit 0? 5 0? 3 0? 2 Time spent playing 2? 3 0 0
Another test given at twelve months
previously separated infants and the controls was one in which an infant was offered vitamins by an experimenter. The previously separated animals were much less willing to approach to obtain the vitamin than were the controls, and this was so even when the test was carried out in the home run. A probable explanation of that is that one of the experimenters had taken part in catching and removing the infants' mothers at the time of the separation.
Eighteen months later when aged thirty months each animal was given a series of comparable tests. On this occasion it was tested over a period of sixteen days when by itself in a screened
-71-
laboratory cage. Out of many tests only a few showed significant differences between groups. One was when the experimenter offered the vitamins. Another was when, on the second and sixth days, a piece of date was hung outside the cage just out of reach: the previously separated animals were much longer than the controls before making an attempt to get it, made fewer attempts, and persisted in making attempts for less long. (The number of animals available for this test made differentiation between once- and twice-separated animals impossible. )
Short-term Effects of a Thirteen-day Separation
At the age of thirty to thirty-two weeks a further six infants were separated for a single period lasting thirteen days (see Spencer-Booth & Hinde 1971b). During the whole of the second week they remained about as depressed and inactive as they were at the end of the first week. (This is in contrast to the moderate degree of recovery seen after the first week of separation in pigtail infants by Kaufman & Rosenblum. )
During the month following the separation it was found that the infants separated for thirteen days were significantly more affected than were infants in either of the other two separated groups. For at least the first week after reunion they emitted more distress calls; and
which showed significant differences between the
66
throughout the month they were more depressed. During the time they spent out of contact with mother they spent more time sitting about inactive than did the other separated infants, and when they were active they were less so. Whereas by the end of the first month of reunion infants separated once and for only six days were showing at least as much activity as they had done before the separation, the activity level of those separated for thirteen days was still significantly reduced. At this time the activity level of the twice-separated infants was intermediate between that of infants separated for a single six-day period and that of those separated for thirteen days.
From all these findings we can conclude with confidence not only that a single separation of no longer than six days at six months of age has perceptible effects two years later on rhesus infants, but that the effects of a separation are proportionate to its length. A thirteen-day separation is worse than a six-day; two six-day separations are worse than a single six-day separation. In these regards the effects of separation from mother can be
-72-
likened to the effects of smoking or of radiation. Although the effects of small doses appear negligible, they are cumulative. The safest dose is a zero dose.
Individual Variations of Response
There is much individual variation in the responses of rhesus infants to separation. Within the age-range studied age had little effect: whether a six-day separation was at twenty-one to twenty-two weeks, at twenty-five to twenty-six weeks, or at thirty to thirty-two weeks seemed to make little difference. Sex played some part: both during separation and after it males were more affected than females. Whether or not an infant was able to cling to another animal during separation had no effect on behaviour after reunion though such clinging did reduce the amount of distress calling at the time.
The most striking results to emerge from the analysis of the data on individual variation are the significant correlations between degree of distress shown by an infant and certain features of the mother-infant relationship ( Hinde & Spencer- Booth 1970). The infants that are most distressed during the first month after separation tend to be those that are most frequently rejected by mother and that play the greatest relative role in maintaining proximity to her. Since in regard to these features there is consistency for each mother-infant pair over time (as measured by rank-order correlations), it is not surprising to find that the degree of distress shown after separation is correlated both with the frequency with which a mother rejects her infant during the period before separation and with the frequency with which she rejects it during the period after reunion. In fact, it is found that, soon after mother returns, correlation of infant distress is higher with frequency of maternal rejection before separation than with frequency of rejection contemporaneously. Subsequently the balance changes and degree of distress becomes more highly correlated with the frequency with which she is rejecting her infant at the time.
Hinde & Spencer-Booth ( 1971) emphasize that these correlations do not justify the conclusion that differences in the mother-infant relation necessarily cause the differences in infants' responses to separations; nevertheless they believe that to be likely.
67
In a recent experiment Hinde & Davies ( 1972) altered the conditions in which separation occurs: instead of mothers being removed from the home cage and placed in a strange one,
-73-
the infants were removed and mothers remained behind. During their thirteen days of separation the behaviour of the five infants conformed to expectations. Although large individual differences made comparisons difficult, the infants separated in a strange cage seemed even more disturbed than did infants that remained in the home cage while mother was removed. After reunion with mother, however, the infants that had been separated in the strange cage were less disturbed than were infants whose mothers had been removed.
Certain observations of the behaviour of the mothers suggest an explanation of this unexpected finding. Compared with mothers that had been removed to the strange cage, mothers that remained behind were less distressed during their infant's absence and, after reunion, were more maternal and less rejecting of it; and harmonious interaction between the two was restored more quickly. These findings tend to support the view that a major determinant of the effect of a separation on a rhesus infant is how mother behaves towards it after reunion.
One form of behaviour that is extremely common in young children after a separation lasting a week or longer in strange surroundings and without substitute mothering, but that has only once been reported for monkey young, is detachment, namely a failure to recognize or respond to mother on reunion. In a study by Abrams (described by Mitchell 1970) twentyfour rhesus infants underwent a two-day separation from mother when aged between eight and twenty weeks. At the time of reunion one-quarter of the infants observed ran away from mother as she approached; and after a second two-day separation a few weeks later, the proportion that ran away doubled. Since, although on the lookout for detachment, Hinde & Spencer-Booth never observed it, the response may prove to be confined to infants in the very young age-range studied by Abrams. It is not yet clear, however, whether the response observed by Abrams can be regarded as homologous with that seen in young children.
The findings of the primate experiments have been described at length because they leave no serious doubt that most of what is to be seen during and after a brief separation in human infants is to be seen also in infants of other species. Explanations of human responses that presume cognitive processes at a specifically human level are thus called in question.
-74-
68
Part II
AN ETHOLOGICAL APPROACH TO HUMAN FEAR
-75-
Chapter 5
Basic Postulates in Theories of Anxiety and Fear
Paradigms provide scientists not only with a map but also with some of the directions essential for map-making. In learning a paradigm the scientist acquires theory, methods and standards together, usually in an inextricable mixture. . . . That is [a] reason why schools guided by different paradigms are always slightly at crowpurposes.
THOMAS S. KUHN ( 1962)
Anxiety allied to fear
Though at intervals down the years one student of the problem after another has been struck that a principal source of anxiety and distress is separation from loved figures, or the threat of separation, there has been great reluctance to accept that simple formula. Objections to it are extremely deep-seated and based on one or more of several common assumptions each of which, it is argued here, is tenable no longer.
In this and the following chapters the simple view is once again advanced. And because there has been so much incredulity and therefore reasoned opposition to it, it is presented in some detail. We begin by considering some of the common assumptions that underlie the traditional incredulity and opposition, with special reference to the influence of Freud's early formulated theory of motivation.
In all psychoanalytic and psychiatric discussion of anxiety it is taken for granted that the emotional states referred to respectively as 'anxiety' and 'fear' are closely related. In just what way they are related is the puzzle. Freud is concerned repeatedly both to compare and to contrast the two: see, for example, Addendum B to Inhibitions, Symptoms and Anxiety ( SE 20: 164168). Others have followed in his steps. Reviewing the whole confused scene in a recent article, Lewis ( 1967) emphasizes that, throughout the broad field of psychopathology, the word 'anxiety' is used habitually to refer to 'an emotional state with
-77-
the subjectively experienced quality of fear or a closely related emotion'. Often, we know, the two words are used interchangeably. In view of the close relatedness of the emotional states concerned and also of the meanings of the two words, it is hardly surprising that ideas about the conditions that give rise to the one state should influence ideas about the conditions that give rise to the other.
69
Nevertheless, in all this diverse, confusing, and contradictory theorizing there is one matter on which all seem agreed: whereas the nature and origin of anxiety are obscure, the nature and origin of fear are simple and readily intelligible.
In the theories advanced in this work there is only one break with that tradition. As hitherto, the states referred to by the words 'anxiety' and 'fear' are seen as closely related. Furthermore, ideas about what arouse states of the one kind are, also as hitherto, closely linked to ideas about what arouse states of the other. Where the paths diverge is at an altogether different point, namely in theories about the nature of the conditions that are apt to arouse what is allegedly the more easily understood of the two kinds of state, namely that of fear itself.
In psychoanalytic and psychiatric circles, it is argued, there still flourish seriously misconceived assumptions about fear and the conditions that arouse it. These mistaken assumptions have long had, and continue to have, a most adverse effect on our ability to understand the distressing anxieties and fears from which our patients suffer.
