No More Learning

-120-
age-groups, however, were not only increasingly more likely to show fear but also increasingly more sensitive to mother's whereabouts. Thus, of those in the two intermediate agegroups (eight and a half and ten and a half months), a quarter withdrew or showed some other sign of fear; and of those aged twelve and a half months no fewer than half withdrew or were otherwise afraid. The effect of mother's whereabouts on the response was only just apparent in the two intermediate agegroups. In infants of twelve and a half months, however, it was abundantly clear. Only when seated on mother's lap did any of these one-year-olds welcome the stranger; when, by contrast, they were seated four feet from her, every one of them showed fear.
Rather similar findings are reported by Bronson (in press) from his short longitudinal study, referred to above (p. 100 ), of infants between three and nine months of age. He observed how the response to a stranger is influenced (a) by the infant's being held by mother and (b) by the infant's being able to see her.
At the age of four months there was little indication that being held by mother reduced wariness when the stranger approached to within two feet and called the baby. But by six and a half months being held by mother reduced wariness considerably and it did so at nine months also.
Presence of an infant's mother within sight, about four feet from him, made little difference to the degree of wariness of the stranger shown by the infant at four and a half or six and a half months. By the age of nine months, however, visual contact with mother was found to reduce wariness. Moreover, at that age it was not uncommon for a baby to crawl over to his mother when the stranger approached.
In the light of these findings it is instructive to consider afresh the much-quoted case of Albert on whom Watson & Rayner reported fifty years ago. In a series of experiments this eleven- month-old infant was conditioned to fear a white rat and, through generalization, a rabbit, a piece of seal fur, and human hair. The unconditioned stimulus was a loud noise made by hitting a long steel bar with a hammer just behind his head. Learning theorists have argued that many cases of phobia are to be traced to conditioning of this kind.
101
Deductions drawn from the case have often been challenged (e. g. see Marks 1969). In the context of this work the following points deserve mention. First, Albert had been 'reared almost from birth in a hospital environment' and was selected for the
-121-
experiment because he seemed so 'stolid and unemotional'. Second, the conditioning took place with Albert placed on a mattress on top of a small table, and with no familiar figure towards whom he could turn. Some of his responses, nevertheless, were those used by a child in turning to a mother figure: for example, raising his arms as if to be picked up and, later, burying his head in the mattress. In addition, Albert was very apt when upset to suck his thumb. To the experimenters this proved most inconvenient, since 'the moment the hand reached the mouth he became impervious to the stimuli producing fear. Again and again . . . we had to remove the thumb from his mouth before the conditioned response could be obtained. ' From these observations the experimenters themselves reached a very significant conclusion: 'the organism . . . apparently from birth, when under the influence of love stimuli, is blocked to all others. '
Thus the results of this early experiment by Watson & Rayner as well as those of the recent ones by Morgan & Ricciuti and by Bronson are consistent with the picture of the growth of attachment behaviour given in the first volume. They are consistent also with two findings by Schaffer described earlier in this one. The first (reported in Chapter 3) is that, before the age of twenty-eight weeks, infants do not protest when removed from mother to the strange surroundings of a hospital but that, from seven months onwards, they do so. The other finding (noted earlier in the present chapter) is that, whereas an infant of twelve months when confronted with strange objects refers constantly to his mother if she is seated behind him, an infant of six months appears oblivious of her being there.
Thus in general it can be said that, just as attachment to a mother figure is becoming steadily better organized during the latter half of the first year, so also is withdrawal from a feararousing situation. Furthermore, because by twelve months a child's cognitive equipment has developed sufficiently for him to be well able to take account of objects and situations briefly absent, he has become able so to organize his behaviour that he moves simultaneously both away from one type of situation and towards another type. Hence he enters his second year equipped to respond in the dual way that is typical of wellorganized fear behaviour. In the next chapter a description is given of how young monkeys pass through the same developmental phases but at a faster rate.
-122-
Fear of Future Contingencies
In this chapter attention has been focused mainly on the nature of the immediate situations that can be observed to give rise to fear behaviour in children. Nevertheless, during the course of human life the situations that are apt to arouse fear include not only those that are actually present but others, more or less likely, that are forecast. Thus children and adults are frequently apprehensive about events that they believe may be going to occur and of objects and creatures that they suspect may be going to appear. Such fear is concerned with future contingencies.
102
Because so many of the situations feared by humans are of that nature and because they bulk so very large in clinical work, it is necessary to examine them in some detail. This is done in Chapters 10 and 11, after the immediate situations that arouse fear behaviour have been considered in the light of the behaviour's biological function.
We turn now to consider what is known about situations that arouse fear in animals. -123-
Chapter 8 Situations that Arouse Fear in Animals
Natural clues to potential danger
Although the stimulus situations that arouse fear in other species are not identical with those that arouse fear in humans, there is much overlap. Overlaps are especially evident, moreover, in the case of the non-human primates, to which much of this chapter is given.
Ethologists take for granted that many of the stimulus situations that arouse fear in animals can be regarded as naturally occurring clues to events that constitute a potential danger to the species in question. This applies especially to situations that arouse fear on the first occasion that an individual encounters them.
Distance receptors are commonly employed for sensing these naturally occurring clues. Depending on the species, an animal may rely mainly on visual clues and receptors, mainly on auditory ones, or mainly on olfactory ones; or on any combination of them. 1 Only when the distance receptors have failed to detect potential dangers in time are the proximal receptors, those for touch and pain, called into action -- and by then it may be too late. Thus, in eliciting fear behaviour, distal clues and distance receptors play a crucial role.
Of all the possible stimulus situations that could act as clues to potential danger and can be sensed at a distance, there are certain ones that are exploited by a very wide array of species. Among the best known are strangeness and sudden approach, both of which regularly evoke fear responses in birds and mammals. Another is the 'visual cliff' to which young mammals of all species so far tested respond by taking avoiding action.
Situations of other kinds, by contrast, arouse fear responses in animals of only a few species; and sometimes, perhaps, of only one. For example, in some species of bird the sight of mammalian fur elicits fear responses; in others, the sight of a pair of staring eyes or of something falling from the sky. In some
____________________
1 For discussions of fear responses in animals see Tinbergen ( 1957), Marler & Hamilton (
1966), and Hinde ( 1970). -124-
species of night-flying moth the high-pitched echolocating calls of predator bats lead to instant flight or, alternatively, to 'catalepsy'. Thus, like drugs, the naturally occurring distal clues to potential danger can be classified into 'broad-spectrum' clues, to which animals of a
103
wide array of species are sensitive, and 'narrow-spectrum' clues, to which animals of only one or a few species are sensitive.
Many of the alarm calls of birds and mammals act as broadspectrum clues since they are responded to with fear not only by members of the species that emits them but by members of other species as well. This is in part because the alarm calls of different species have come to resemble each other, presumably through a process of natural selection.
In a number of animal species olfactory stimuli, some of broad but many of narrow spectrum, are especially effective in eliciting fear behaviour. Such 'warning scents' arise from one of two sources: from enemies or from friends. On the one hand, as is well known, the scent of an approaching predator, man or wolf, can elicit fear responses in a broad array of grazing mammals, zebras, deer, and antelope. On the other, an 'alarm scent' emitted by an animal when frightened or wounded can elicit fear responses in other animals (exactly as an alarm call can), but in this case the effect is likely to be confined to members of its own species.
Thus animals of every species are born genetically biased so to develop that they respond with one or another form of fear behaviour whenever they sense a stimulus situation that serves as a naturally occurring clue to one of the particular dangers that beset members of their species. Since some categories of potential danger are common for a wide array of species, clues to them act as broad-spectrum clues. Since other potential dangers affect only a few species, clues to them are likely to be narrow-spectrum.
Just as in man the forms of behaviour that can conveniently be labelled as fear behaviour are diverse, so are they in nonhuman species. Responses include, on the one hand, crouching, curling up, freezing and taking cover, and, on the other, calling, escaping, and seeking proximity to companions. The precise response shown turns on many factors -- the animal's species, its sex and age, its physiological condition, and also the particular type of situation that has aroused the fear.
For example, Hinde ( 1970) reports a finding by Hogan that, in chicks, withdrawal occurs from stimuli at high intensity (and
-125-
some others) whereas freezing is elicited by stimuli that are strange, novel, or surprising. Again, both Lorenz ( 1937) and Tinbergen ( 1957) have pointed out how, in many species of bird, distinctive situations can elicit distinctive sorts of response. The Burmese jungle fowl (and also the domestic chicken) possess two distinct warning calls uttered in response to the sight, respectively, of a flying raptor and a terrestrial predator. When heard by another fowl, the raptor-type warning call elicits a downwards escape, ending, where possible, beneath cover of some kind. When, by contrast, the predator-type warning call is heard a fowl takes off and flies into a tree. These distinctive types of behaviour in response to distinctive warning calls are further evidence that, as is indicated in Chapter 6, we are dealing, not with some single and comprehensive 'instinct of fear', but with a heterogeneous collection of interrelated forms of behaviour, each elicited by a slightly different set of causal conditions.
Fear behaviour, it has been emphasized, may not only remove an animal from situations of certain kinds but take it towards, or into, situations of other kinds. Depending on the warning call heard, a fowl flies into ground cover or into a tree. A form of behaviour shown by
104
animals of a great many species, and one of especial interest to our thesis, is movement that takes an animal towards his companions. For example, when a peregrine is overhead, lapwings not only take flight but keep close together as a flock; starlings do the same. (By contrast, in the same situation partridges crouch close to the ground. ) Most of the group-living mammals also edge closer together when alarmed. Movement of this kind is particularly evident in young mammals which, with only few exceptions, habitually run to mother and stay close to her.
Let us return now to the situations that arouse fear. It is probable that all the examples of distant situations mentioned so far in this chapter are responded to by fear behaviour of one kind or another on the very first occasion that an individual of a particular species encounters them. In such cases no special opportunities for learning that the situation is potentially dangerous are required. In the case of other stimulus situations, however, the position is quite different. Only after the situation has become associated with some other clue to potential danger is a fear response aroused. A clue universally known to lead to such learnt associations, though not the only one, is pain.
-126-
Pain receptors are proximal and therefore their role is in many ways different from that of the distal receptors. In the first place, pain receptors are usually called into action as a last resort and only when the distal receptors, or the fear responses they may have elicited, have failed to ensure the animal's withdrawal. In the second place, the sensation of pain leads commonly to immediate and urgent action. In the third place, the sensation of pain may well mean that the danger has already materialized. For these reasons it is easy to suppose that pain and danger are in some way identical, which of course is not the case (see next chapter), and thus to give pain far too great a prominence in theories of fear behaviour.
Because by being a proximal clue to potential danger pain is very late in acting, it is of great biological advantage to an animal to learn to recognize potentially painful situations from associated distal clues. The investigation of such learning has for long been a principal interest of experimental psychologists, and in consequence much is known about it. In particular, it has long been known from conditioning experiments in which a neutral stimulus is coupled with a painful one that, in a great variety of mammalian species, a fear response to a stimulus hitherto neutral is both quickly established and very hard to extinguish.
A high concentration of interest on the fear-arousing properties of pain, and on the learning to which it gives rise, has at times led to a neglect of the immensely important and prior role of distal clues and distal receptors, both in animals and in man. As a result it is not always realized that in many species a new distal clue to potential danger can be learnt as readily by watching how companions respond to it, and then copying them, as by its becoming associated with pain. In mammals, indeed, a principal means whereby new situations come to be categorized as potentially dangerous, and so to be responded to by fear behaviour, is that of copying older animals, especially parents. In no kinds of mammal does imitative behaviour of this sort play so large a part as it does in primates.
Fear behaviour of non-human primates
Some years ago, as a result of long experience with chimpanzees in captivity, Yerkes & Yerkes ( 1936) wrote: 'The stimulus characters which early and late are dominant in the
105
determination of avoidance responses are: visual movement, intensity, abruptness, suddenness and rapidity of change in stimulus or
-127-
stimulus complex. ' Although this description needs a little elaboration the nub of the matter is there.
Field Observations
Field observers of the primates are well aware that sudden noise or sudden movement is immediately effective in alarming their subjects and leading to their rapid disappearance. Describing her experiences in watching langur monkeys in the forests of India, Jay ( 1965) writes: 'Forest groups gradually became used to me and I could follow them at a distance of about 50 feet. However, if any sudden movement in the brush startled them, they immediately fled from sight. ' Sudden sounds have the same effect.
For a forest-dwelling species, such as the langur tends to be, safety lies anywhere in the tree- tops. For ground-living species, by contrast, safety may lie in only one special place. For example, in East Africa the home range of each band of olive baboons must contain at least one clump of tall trees to the tops of which the band retreats whenever alarmed and in which it sleeps ( DeVore & Hall 1965). Further north, in Ethiopia, family parties of the related species of Hamadryas baboon must live within reach of precipitous cliffs to which they similarly can retreat ( Kummer 1967). The location of their haven of safety is a supreme determinant of the behaviour of these animals: 'Where large predators such as lions are numerous . . . the absence of trees in some areas may deny baboons access to rich food sources when food items in general are scarce' ( DeVore & Hall 1965).
In the field studies of non-human primates so far published, systematic attention is not always given either to the situations that evoke fear behaviour or to the forms that the behaviour commonly takes. The long-term study of wild chimpanzees undertaken in Tanzania by van Lawick-Goodall ( 1968) presents far more detail than most.
Van Lawick-Goodall starts by emphasizing that the form that fear behaviour takes 'depends on the situation and the individual or individuals concerned'. When a chimpanzee is startled by a sudden noise or movement nearby, its immediate response is to duck its head and to fling one or both arms across its face; alternatively, it may throw both hands in the air. Occasionally these startle reactions are followed by a hittingaway movement with the back of the hand towards the object, at other times by flight. When the alarming object is another
-128-
and more dominant chimpanzee, flight is accompanied by loud screaming; when it is anything else, flight is quite silent. An alternative to flight is cautious withdrawal out of sight, combined with an occasional peering-out to see what is going on.
The situations van Lawick-Goodall reports as having evoked startle responses involved sudden noise or movement, for example a low-flying bird, a large insect, or a snake. Fear responses were very often aroused in a chimpanzee when another, more dominant, animal was making threatening gestures. Before the chimpanzees got used to her presence the observer
106
herself was a common source of both fear and flight. After a year or so many of them carried on their normal activities when she was as close as thirty to fifty feet from them. Nevertheless, they quickly became uneasy if she began to follow them; often, too, she had to conceal her interest in them by diversionary activity, such as pretending to eat leaves or to dig.
Since so many species of animal give an alarm call when frightened, van Lawick-Goodall was surprised that the chimpanzees she studied never did so (except when fleeing from one of their own kind). Instead, each one moved away silently on its own. Nevertheless, they were quick to be alerted by the alarm calls of other species: they 'were invariably alerted by the alarm barks of baboons and also by the alarm calls of other monkeys, of bush buck and of some species of birds; after hearing such calls, they peered round to ascertain the nature of the disturbance'.
As in so many other species, to move away from an alarming situation or event is, in chimpanzees, only half the picture of fear behaviour. The other half is to move towards some place treated as though safe or to make physical contact with companions. The latter was seen in adults as well as in young. Van LawickGoodall describes how adult animals when frightened move towards and hug one another. This behaviour she believes to be a direct extension of what is seen so regularly in the infant:
Thus a mature chimpanzee may embrace, reach out to touch, or mount another animal under similar circumstances and in more or less the same manner as a frightened or apprehensive infant runs to embrace or be embraced by its mother, reaches out to grasp or touch her hair, or stands upright behind her grasping her rump . . . ready to climb on if the situation warrants it.
The calming and reassuring effects of contact with another
-129-
animal are discussed by van Lawick-Goodall in some detail. A touch, a pat, or an embrace from a dominant animal was quickly effective in calming a subordinate, and occasionally the reverse occurred. One mature male was seen to find comfort by embracing a female only three years old, once when he had a sudden fright from seeing his own reflection in a glass and twice after he had been attacked by another male.
Field observers of other primate species have also noted the strong propensity of a frightened or agitated animal to touch or cling to a companion. For example, in his description of the behaviour of wild Hamadryas baboons, which live in stable family units of one male with up to three females and their young, Kummer ( 1967) remarks that not only infants but adults also when under stress are strongly disposed to cling to a companion. Thus an adult female, when alarmed, clings to the back of her husband or is embraced by him. Conversely, when he is under stress during a fight, a male is likely to embrace one of his wives. When an animal that has left its mother but is still not fully mature becomes frightened it seeks out the highest- ranking individual within range. Since not infrequently it is the threats of this animal that aroused the fear in the younger animal the result is paradoxical: the young animal runs to and clings to the very individual that aroused its fear. Among many interesting features of Kummer's study is the evidence he presents that in this species the relationship of a mated male and female is patterned closely on the relationship of a mother and her infant.
107
The persistence into adult life of patterns of behaviour seen first and at greatest intensity during infancy is found, then, to be a regular feature of the behavioural repertoire of other primate species. It warns us against supposing that, whenever something similar is observed in humans, as it so often is, it must be treated as an example of regression.
In wild animals it is never possible to be sure whether an individual would respond to a particular situation with fear were it to encounter it for the first time or whether it does so only after having learnt to do so. Fear of snakes is a case in point. Van Lawick-Goodall reports that the wild chimpanzees she observed showed fear both of a fast-moving snake and of a dying python. Yet it seems that chimpanzees brought up in zoos do not always show such fear. 1
____________________
1 A great many zoologists, including Charles Darwin, have been interested in the marked
tendency for monkeys and apes to respond to -130-
Apparently incompatible findings of this kind are not difficult to reconcile. In a social species, to respond with fear to a situation, once learnt, is handed on by tradition. This point is well illustrated by an observation made in Nairobi Park ( Washburn & Hamburg 1965). A large band of some eighty olive baboons were sufficiently tame to be approached easily in a car. Two of these baboons were then shot (by a local parasitologist). Thenceforward the baboons fled on sight of man or car, and eight months later they still could not be approached although they had seen 'harmless' cars almost daily during the interval. This example is in keeping with the common finding that a response learnt as a result of a single violent experience does not extinguish quickly. It illustrates, further, that it is not necessary for more than a few animals in a band actually to have been exposed to the alarming experience, since it is customary for all the animals in a band to flee as soon as they either hear an alarm bark or see a dominant animal running off. Thus, by following a tradition once set by their elders, members of a band may for years treat whatever happens to have frightened one of their number, present or past, as potentially dangerous. By these means, a tradition that snakes, or men, or cars are to be avoided may develop and persist in one social group though not in another.
Until recent years there was a tendency to suppose that maintenance within a social group from generation to generation of special ways of behaving was a skill confined to man. Now it is recognized that cultural traditions occur also in many other species and affect many forms of behaviour: how to sing ( Thorpe 1956), what to eat ( Kawamura 1963), where to nest ( Wynne-Edwards 1962). It is no surprise, then, to find that in a bird or mammal species cultural traditions exist regarding what to avoid.
The part played in human development by culturally determined clues to potential danger is discussed further in Chapter 10. Here it may be noted that recent experimental studies of monkeys demonstrate clearly that an animal may
____________________
snakes by strong fear, often amounting to panic, and many observations are on record. The evidence is reviewed by Morris & Morris ( 1965), who also record striking observations of their own. While some measure of learning cannot be ruled out, it is evident that in old- world monkeys and apes the tendency to fear snakes is very pronounced, is relatively specific and, if learnt, is remarkably long-lived in the absence of any further experience.
108
-131-
learn to fear a situation solely by observing how a companion responds. For example, Bandura ( 1968) refers to a study by Crooks which shows that monkeys that initially played freely with certain play objects ceased to do so after they had witnessed another monkey (apparently) emit cries of fear whenever it touched one of the objects. 1
Experimental Studies
Many other studies of captive animals, including experimental studies, fill out our knowledge of the fear behaviour of nonhuman primates and of the situations that are likely to evoke it.
Two visual situations that arouse fear in young rhesus monkeys are a looming stimulus and the visual cliff. Both experimental situations are described in the previous chapter where the fear responses of human infants are discussed.
Schiff, Caviness & Gibson ( 1962) studied the behaviour of twenty-three rhesus monkeys of varying age when confronted by a looming stimulus; eight were infants of between five and eight months, and the remainder adolescent or adult. Each animal was tested alone in its own cage at a distance of five feet from the screen on to which the expanding (looming) shadow was projected. All but four of the animals responded immediately by either withdrawing or ducking. A number of animals sprang to the rear of the cage, often bumping hard against the back. Other and less active animals were quick to duck head and upper part of body. Younger animals often gave alarm calls as well. (The four animals that failed to respond were thought to be looking elsewhere when the stimulus was presented. ) No age differences were shown. The speed and form of the stimulus appeared irrelevant. No habituation occurred when two animals were each exposed to a series of fifteen looming trials at intervals of ten seconds.
When the same animals were confronted by a contracting (receding) shadow the response was quite different. All but four remained at the front of the cage and appeared interested as the shadow contracted. A general brightening of the screen also aroused interest. A darkening screen produced no particular response, except when it was presented after a looming stimulus: then it produced a few slight flinches, much milder than those that occurred to looming.
____________________
1 In fact the distress vocalizations were played on a tape-recorder each time the monkey
touched an object. -132-
The number of young rhesus monkeys tested on the visual cliff is few, but responses are unambiguous. Walk & Gibson ( 1961) report on a male infant tested at ten days, and again at eighteen and forty-five days, and on a female infant tested at twelve and thirty-five days.
During their second week both infants proved only fairly reliable in avoiding the 'chasm'. By the ages of eighteen and thirty-five days, respectively, the two showed pronounced discrimination and effectively avoided the 'deep' side at every test. Thus, in this species, avoidance of the deep side is only partly efficient when locomotion begins, but it improves rapidly. The results of similar experiments on another small sample of rhesus infants, reported by Fantz ( 1965), are of a similar kind.
109
Strangeness has been used as a fear-arousing stimulus in many experiments with primates.
Harlow and his colleagues have conducted a number of experiments on the fear behaviour of young rhesus monkeys. 1 Before about twenty days of age an infant rhesus shows no sign of fear of strange visual stimuli; for example, it will confidently approach a moving toy animal it has never seen before. After that age, however, and especially after six weeks, the presence of such a toy leads an infant immediately to rush away from it. Infants that have been reared on a cloth dummy 'mother' not only flee from the alarming toy but return promptly to the familiar dummy mother, to which they then cling tightly. Often a rather older infant, of twelve weeks or more, having fled from the alarming toy and clung tightly to its familiar dummy mother, relaxes. Then it may leave the dummy mother and cautiously approach the fear-inducing toy; it may even explore it manually. The behaviour of the same infant when its familiar dummy mother is absent is, however, very different. It is likely then to curl up on the floor and scream (see below, p. 135 ).
Mason ( 1965) has carried out rather similar experiments with chimpanzees, also using strangeness as a main form of feararousing stimulus situation. In this species also behaviour is very different according to whether an animal is with others or alone. This leads to a consideration of the effects on nonhuman primates of compound situations, and especially of the striking effects of being alone.
____________________
1 An account of some of Harlow's experiments is given in Volume I, Chapter 12. See also
Harlow & Zimmermann ( 1959); Harlow ( 1961); Harlow & Harlow ( 1965). -133-
Compound situations
Monkeys and apes are like humans in that, when confronted by a situation compounded of more than one alarming feature, they are apt to exhibit fear at an intensity far greater than they would were any one of the features to be present singly. Being alone in the presence of a fear- inducing stimulus, moreover, greatly intensifies the fear behaviour seen.
Being Alone
An experimental study reported by Rowell & Hinde ( 1963) gives quantitative data for a sample of seventeen rhesus monkeys, thirteen adult (three male and ten female) and four subadult (two of each sex). These animals live together in stable groups of a male with three or four females and young. The tests, each of which lasted for three minutes, consisted of very simple situations. In each test the tester, who was well known to the animals, stood close to the cage. In one he offered them pieces of banana; in another he stood quietly watching but not staring; and in a third he dressed up in mask and cloak and made slight movements. Before being tested each animal was observed for half an hour and its behaviour was recorded. Thereafter the three tests followed, separated by intervals of five minutes.
In the first series of tests the animals were tested while living together in their regular groups. Each time the tester appeared they showed a characteristic change in behaviour compared with what they had shown before the testing started. There was a great increase of threat noises and of activities, such as lipsmacking, scratching, and yawning, that are associated with
110
stress. In addition, they urinated more frequently, their hair stood on end, and they showed a frightened facial expression. (Attacks were sometimes made by the adult males towards the tester but were not made by the other monkeys. )
Most of these forms of behaviour were considerably more in evidence when the observer wore the mask and cloak and moved than when he stood quietly by. Of responses seen in the mask test, significant increases in frequency were recorded for low aggression threat noises, hair standing on end, urinating, frightened expression, and yawning. In general terms it seems that, whereas the monkeys were merely 'made uneasy' when the observer quietly watched, when he wore the mask they became 'alarmed and angry'.
In a second series of tests each animal was tested alone. For
-134-
a period of six hours before testing began the remainder of its group was locked into an indoor cage while the animal under test remained alone in its familiar outdoor cage; it could, however, hear its companions and see them through a window, so that it was far from being isolated. Nevertheless, for every animal, fear responses to the simple tests were far more frequent when it was alone than when it was with its group. Increased scores ranged from threefold to fiftyfold. The response showing the greatest increase in frequency was that of looking into the window where it could see its absent companions.
Summing up their findings, Rowell & Hinde write:
Thus isolation is best regarded not merely as an additional stress-producing factor acting equally under all circumstances, but rather as one which, while producing relatively little effect on undisturbed animals, can strongly accentuate the effect of other stress-producing agents. It is as though isolation multiplied their effects, rather than summating with them.
The results of Harlow's experiments on young rhesus monkeys brought up on dummy mothers strongly support this conclusion ( Harlow & Harlow 1965). In one series of experiments four infants raised on cloth dummy mothers were introduced, singly, to a strange 'room', six feet square, containing various objects known to be of interest to young monkeys. Each week two tests were given to each infant. In one the infant's dummy mother was present in the 'room', in the other it was absent. According to whether its dummy mother was present or absent the behaviour of an infant was utterly different.
When the dummy mother was present an infant, on entry to the strange room, rushed to it and clutched it tenaciously. The infant then relaxed and, showing little sign of apprehension, began climbing over the dummy mother and manipulating it. After several such sessions, the infants began to use the dummy mother as a base from which to explore. Leaving the dummy, an infant moved to a toy, picked it up and handled it, and then moved back to the dummy. Sometimes a monkey would bring its plaything with it. Exploration of an object away from the dummy mother alternated with rapid return to base. Throughout, the monkey seemed relaxed and confident.
In the absence of the familiar dummy mother, behaviour was radically changed. An infant either curled up on the floor rocking and crying, or else ran around clutching hold of itself.
111
-135-
Exploration of the objects, when it occurred at all, was 'brief, erratic and frantic'. The impression given to an observer was of an infant in a state of distress and misery. 1
The results of Mason's experiments with young chimpanzees point in the same direction ( Mason 1965). In one experiment twelve African-born animals were, singly, given electric shocks to the foot, both when held by an observer and when alone. Whereas when an animal was alone it whimpered and screamed some 60 per cent of the time, when held by an observer it was practically silent. Comparable results were obtained when the animals, instead of being shocked, were confronted by a novel situation.
In yet another series of experiments, conducted by Gantt in a Pavlovian tradition, it is shown that anxiety, induced experimentally in dogs, is much reduced by the presence of a human companion, especially someone well known to the animal. Patting and petting the dog is particularly efficacious; and the effect is more pronounced on animals made 'neurotic' by frequent experimental procedures than on more normal animals. Findings are reviewed by Lynch ( 1970).
Fear, attack, and exploration
Stimulus situations that are likely to arouse fear in humans and other animals can also, when circumstances are a little changed, evoke behaviour of quite different sorts. Attack is one of these alternative forms of behaviour; exploration another.
Whether an animal flees from a potentially fear-inducing stimulus object or goes in to attack it turns on very many factors, some organismic, others situational. Of the organismic factors, the individual's species, age, and sex play major parts. In many species, including the ground- living primates, older animals, especially males, are more likely to attack, whereas immature animals and females are more likely to withdraw. Ill health and fatigue may also play a part in tipping the balance towards withdrawal. Hunger often tips it towards attack. Of the situational factors, being on familiar territory makes for
____________________
1 Infants that had been reared on a wire dummy mother were little influenced by whether it
was present or absent. In both conditions they showed very distressed behaviour, which was at a level of intensity significantly greater even than that of the infants reared on a cloth dummy mother when the dummy was temporarily absent. Thus the wire dummy mother proved entirely ineffective as a base from which to explore.
-136-
boldness, being elsewhere for withdrawal. When escape routes are blocked, attack is the rule. Not infrequently, behaviour of both sorts is clearly aroused: even in the act of attacking, an individual may show signs of also being afraid. Because of the close association between them, attack, threat, flight, and submission are sometimes lumped together by ethologists and termed 'agonistic behaviour'. The reason for the close association between these forms of behaviour is that, of the many causal conditions necessary to elicit each one of them, some are shared in common ( Hinde 1970).
112
This circumstance is the explanation also of the close link there is between withdrawal and exploration, discussed in Chapter 13 of Volume I . It is well known that a single kind of stimulus situation, namely strangeness or novelty, can elicit either withdrawal or exploration or both together. In animals of many species a small change in the environment elicits investigation whereas a larger one arouses fear behaviour. Not infrequently an interested approach and an alarmed withdrawal are shown either simultaneously or in rapid succession. Which of the two classes of behaviour becomes dominant turns on many factors -- the details of the novel stimulus, the environment in which it is met (familiar or unfamiliar terrain, companions present or absent), the age and sex of the individual, its hormonal condition, and no doubt other factors besides.
The fact that small changes in a situation can have great influence on the form of behaviour shown cannot be overemphasized. If a population of animals is to survive in the wild each one needs to show, according to its age, sex, and social status, a nice balance between discretion and valour.
-137-
Chapter 9
Natural Clues to Danger and Safety
I left my darling lying here, a-lying here, a-lying here,
I left my darling lying here, To go and gather blaeberries
I found the wee brown otter's track, The otter's track, the otter's track,
I found the wee brown otter's track, But ne'er a trace of baby-O
I found the track of the swan on the lake, The swan on the lake, the swan on the lake, I found the track of the swan on the lake, But not the track of baby-O
I found the trail of the mountain mist, The mountain mist, the mountain mist, I found the trail of the mountain mist, But ne'er a trace of baby-O
From the Gaelic
Better safe than sorry
None of the stimulus situations so far considered -- strangeness, sudden change of stimulation, rapid approach, height, being alone -- is intrinsically dangerous. Each one is no more than an indicator of potential danger or, more precisely, of an increased risk of danger, and, as such, of only moderate accuracy. As a result, much fear is aroused in situations that later turn out
113
not to have been dangerous at all; while, by contrast, certain truly dangerous objects and events are heralded by no fear-arousing natural clues. This imperfect correlation of natural clues with actual dangers has proved confusing for clinicians and a trap for unwary theorists.
The heart of the theory here advanced, which derives directly from ethology, is that each of the stimulus situations that man is genetically biased to respond to with fear has the same status as a red traffic light or an air-raid siren. Each is a signal of potential danger; none is intrinsically dangerous. In a similar
-138-
way, each of the stimulus situations that man, when alarmed, is genetically biased to approach and cling to has the same status as sanctuary on sacred ground. Each signifies potential safety; none is intrinsically safe. Whereas the signal value of the red light and the sacred ground is conferred by human convention and transmitted by word of mouth, that of the natural clues is conferred by statistical association and transmitted by genes. Strong genetically derived biases to respond differentially to these two classes of natural clue either by withdrawal or by approach have, during the course of evolution, become a characteristic of the human species because of their survival value. Most clearly apparent during childhood and old age, sometimes disguised or discounted during adult life, these biases nevertheless remain with us. From the cradle to the grave they are an intrinsic part of human nature.
This theory, it will be seen, explains well why, in modern Western environments, fear can be readily aroused in situations that are not, in fact, the least dangerous; and also why fear can be readily allayed by actions, such as clutching a teddy bear or sucking a pipe, that do nothing effective to increase safety. Though to the eye of an intellectual city-dweller such behaviour may seem irrational and childish, and may even be attributed to pathological fancy, to the eye of a biologist a deeper wisdom is apparent. Examination shows, indeed, that, so far from being irrational or foolhardy, to rely initially on the naturally occurring clues to danger and safety is to rely on a system that has been both sensible and efficient over millions of years.
For, it must be remembered, we have but one life. Though on occasion risks are run, either for potential gain or merely for fun, in the ordinary run of days it is better by far when natural clues are perceived to take what on ninety-nine occasions proves to be unnecessary action than, by habitually ignoring them, to fall victim on the hundredth. Were we regularly to ignore red traffic lights we might not meet our doom for a time; but our days would be numbered.
A natural clue to potential danger signals merely an increased risk of danger and gives no information regarding the absolute degree of risk. For animals of different species, of different ages and sex and in different environments, the absolute risk indicated by one or another clue can vary from high to quite low. For example, certain natural clues that are closely linked to predators, such as staring eyes, may perhaps in certain natural environments be associated with a very high degree of
-139-
risk, whereas in some other environments the risk might be low. Similarly, certain other natural clues, such as being alone, might be associated with either a high or a low degree of risk depending on the particular circumstances and the particular individual. Nevertheless,
114
whatever the absolute levels of risk may be, a natural clue is associated as a rule with a raised degree of risk. The increase may be from moderate to very high or perhaps from near zero to a mere 1 per cent. Without a great deal of knowledge of the total situation the absolute degree of risk in any one case cannot be known. What seems clear, however, is that in every case the degree of risk is likely to be raised.
The great advantage of our being biased to respond, by prompt withdrawal, to the natural clues to an increased risk of danger is that singly, and especially together, they act as indicators to a high proportion of all the dangerous situations into which we might stray. No matter that they embrace also a great many situations that are not dangerous at all. Far better to be safe than sorry. ?
In an analogous way it is also an advantage when running away from a potential danger to run towards a potential haven of safety: for small animals ground cover, for monkeys the tree- tops, for group-living species the social band, for weaker animals their stronger companions. No matter if such action is taken unnecessarily: once again, better safe than sorry.
By this point some readers may have become impatient. However true the principles outlined may be for monkeys and apes, and perhaps even for young human children, adult humans are a great deal more sophisticated than to attend merely to natural clues. Thought and imagination, rational or irrational, conscious or unconscious, are the stuff of human fear. Why waste time on these primitive mechanisms? The reason is, of course, that much of the sophisticated superstructure of cognitive and feeling processes characteristic of Western man in the realm of fear is intelligible only in terms of the primitive genetically biased groundwork that evolved in a different environment and that we share with other primate species. A failure to understand this primitive groundwork, it is argued, has led to many and serious misunderstandings. Not only is the behaviour of every human adult influenced by these primitive processes but so also are his most sophisticated cognitive structures and his most sensitive ways of feeling.
-140-
Whether suddenly alarmed or chronically anxious, whether temporarily comforted or steadily confident, the way a man or woman thinks and feels is determined in significant degree by these strong genetic biases to respond unthinkingly to the natural clues.
In the chapters that follow attention is given, first, to showing that the strong tendency to respond to the natural clues accounts for most of the more elaborate situations that humans come to fear, and, subsequently, to the way in which increasingly refined processes of appraisal lead to a broad spectrum of human feeling states. Before proceeding, however, let us consider further the genetically biased groundwork. We begin with the special place of physical pain as a natural clue.
The Limitations of Pain as a Natural Clue
In the past there have been theorists who have postulated that almost the only type of stimulus to which there is any genetic bias to respond with fear is physical pain, and that all other stimuli derive their fear-arousing properties from becoming associated with pain. Not only is the theory false, but a moment's thought shows it to be hardly plausible.
115
As a natural clue to potential danger, the experience of physical pain is in a special category. The clues to which attention has so far been directed are distal clues perceived by the distance receptors, eye, ear, and nose. By giving warnings while potential danger is still more or less remote, these clues enable an animal or man to take precautions in good time. By contrast, as noted in the previous chapter, to await events until pain is experienced may well be to wait too long. Whereas the distance receptors can be likened to advance look-outs, physical pain has the status of last ditch.
The special property of pain is that, being so late in acting, it leads to immediate and urgent action. The phase of alert wariness, so characteristic of many animals after a distal clue is first sensed, is absent. Instead, there is immediate and unthinking retreat, or, alternatively, attack.
Another special property of pain is, of course, its power to promote learning. Countless experiments demonstrate how rapidly and firmly an animal learns to recognize a situation in which it has experienced pain and to respond thenceforth by avoiding it. After such learning, an animal no longer relies on the hazardous proximal clue of pain but comes instead to use some distal clue that gives time and space in which it can take
-141-
precautions. The advance look-outs are alerted to identify and beware of a new clue.
Even though physical pain may be more highly correlated with potential danger than are some of the other natural clues, it is not infallible. For example, medical attention may be painful but is usually not dangerous; whereas a truly dangerous condition, such as internal haemorrhage, may be accompanied by no pain. That is but one example of a serious danger that is either without natural clues or heralded by faint ones only.
Dangers that have no Natural Clues
Earlier it was noted that the natural clues to which we react with fear are, singly and especially together, indicators of a high proportion of all the dangerous situations into which we might stray. Nevertheless, there are some dangerous situations that present no clue to which we have a natural bias to respond by escape. Some, indeed, even emit no signal that our sense organs can detect.
Among naturally occurring hazards, infectious illnesses are cases in point. Where infection is airborne there is usually no naturally occurring clue that we are able to sense and from which we are genetically biased to withdraw. (In contrast, by producing bad smells or tastes, food- and water-borne infections are often much less silent. ) In modern times, moreover, man has added a number of other dangers that also emit no clues to which human nature is sensitive. Examples are carbon monoxide gas and X-rays. Since in such cases evolution has as yet had neither time nor opportunity to provide us with natural means for their detection, we have to rely instead on man-made indicators.
Thus, although by exploiting the natural clues to danger and safety our genetic endowment provides us with a remarkably sensitive and efficient means of protection, it is far from foolproof. On countless occasions we are led unnecessarily to avoid wholly harmless situations; on a few others we are permitted to blunder into truly dangerous ones.
116
Potential danger of being alone
The natural clue to increased risk of danger with which this volume is especially concerned is being alone. Statistically, being alone is less safe than being with a companion. That that should be so during childhood, during sickness, and in old age
-142-
may not be too difficult to grasp. That it is also the case for the ordinary healthy grown-up man and woman may at first sight seem unexpected. Yet there is good reason to believe that it is so, especially in certain situations; even though in Western countries the situations may be few and the absolute risk not high. The thesis of this section is, therefore, that, as it was throughout man's earlier history, in many circumstances still today it is as appropriate to avoid being alone as it is to avoid any of the other natural clues to potential danger. That we should be so constructed that we find comfort in companionship and seek it, and that we experience greater or less degrees of anxiety when alone, is, therefore, in no way surprising.
In the previous volume (Chapter 4) it is argued that, if we are to understand human behavioural equipment, it is necessary to view it in the light of what we know of man's environment of evolutionary adaptedness. Later, following this line of thought (Chapter 12), the theory is advanced that in man's environment of evolutionary adaptedness the function of attachment behaviour, which of course promotes proximity to special companions, is protection from predators, and that this is as true for humans as it is for other species of mammal and bird. For all ground-living primates, safety lies in being with the band. To become separated from it is to provide a more or less easy meal for a lurking leopard 1 or a pack of hunting dogs. For weaker members, especially females and young, the old and the sick, isolation often spells speedy death.
By practical people this theory is sometimes treated as an academic curiosity. Yes, it may be said, there may well have been a time in man's history when separation entailed danger from predators. But that was long ago. For such responses to persist into modern times is an irrelevant nuisance. It is time to rid ourselves of such archaic superstition.
This line of reasoning has several defects. In the first place, even if we wished it, genetic biases built in over millions of years cannot be eradicated overnight. In the second, reflection suggests that to try to eradicate them might be most unwise. For in many parts of the world today the absolute risk attendant on
____________________
1 Since publication of the first volume further evidence has come to light of the dangers of
leopards to early man. According to Brain ( 1970) the fossilized bones of Paranthropus robustus found in a cave in the Transvaal are fragmented in ways typical of leopard prey. One of the betterpreserved skulls (of a juvenile) bears two holes the right size and distance apart to fit the canine teeth of a leopard.
-143-
being alone is still fairly high; and even in Western societies the risk may be higher than we like to imagine.
117
In Western countries today, it is true, injury and death are no longer due to predators. But there are other dangers. In place of predators, power-driven motor cars and household equipment provide novel hazards and take their toll. Young children newly mobile and the elderly are among the principal victims. Yet, though ordinary experience strongly suggests that those most at risk are children and old people left on their own, researchers concerned with accident prevention seem to have given the matter little attention. Statistics of traffic accidents for one of the London boroughs and also for Sweden are, however, revealing.
Traffic Accidents to Children
During 1968 in the London borough of Southwark, 1 injuries to pedestrians numbered 901, of which twenty-seven were fatal. Of the total injuries, 411, or nearly half (46 per cent), were to children under the age of fifteen. This represents an incidence of injury to children about three times that to adults.
The most vulnerable age-group were children between their fourth and eighth birthdays. At these ages the risk of injury was about five times that for adults. The incidence for those a little younger and a little older (the three- and the eight-year-olds) was only slightly lower. The age distribution was as shown in the table below.
Age in No. of
years
0- 2? 11 3- 5? 11 6- 8? 11 9-11?