The Act of Creation

by Arthur Koestler

  Another distortion of instinct-behaviour are the so-called "displacement activities" which overlap with Leerlauf (and also with play). 'Displacement is the performance of a behaviour pattern out of the particular functional context of behaviour to which it is normally related. It seems to appear when the charge (SAP) attached to one instinct is denied opportunity for adequate discharge through its own consummatory act or acts and instead sparks over to set going the comummatory act of another instinct'. [28] A dog in its restraining harness in a Pavlov-type laboratory, while expecting the fall of food from the container, will stamp, yawn, and pant -- activities which do not belong to his normal feeding behaviour; but what else, one may irreverently ask, can the poor excited creature do? Pail-fed calves will suck the ears or navels of their companions, as infants suck their thumbs. Some birds play elaborate games, throwing up and catching sticks; so do puppies; kittens will 'pretend' that a ball of wool is a mouse. Leerlauf and 'displacement' thus comprise a broad range of activities which occur in the absence of the proper stimulus or in the presence of normally inadequate ersatz stimuli; or when the proper response is for some reason blocked. Its human equivalents range from playful activities to repetition compulsions and the formation of neurotic ersatz symptoms.

  Instinct and Originality

  At the opposite end of the rigidity-flexibility scale we find adaptations of instinct-based behaviour-patterns which give the impression of original improvisation. Even the ritual-bound stickleback, that stickler for etiquette, is capable of them: 'If the normal behaviour-pattern is continually interfered with, quite large modifications in the normal instinctive orientation of the nest-building movements may be made.' [29]

  Thus the denial of normal outlets can lead either to the mechanical reeling-off of the built-in pattern in freewheeling or displacement activity; or to original re-adaptations of the pattern. Which of these alternative possibilities will occur depends on the nature of the challenge, and the animal's 'ripeness' to cope with it. What solution, after all, could even a genius cat find to comply with its code of hygiene on the kitchen tiles? What creative outlet is left to the squirrel to solve his nut-hiding problem?

  On the other hand, ethologists have produced many striking examples of ingenious instinct-based behaviour in the face of adversity. The female of a certain wasp, Eumenes conica, builds clusters of clay-cells or pots, deposits an egg in each, provisions it with caterpillars for food, then closes the cells with clay lids. If now an artificial hole is made in a cell, the wasp will first stuff the caterpillars which have fallen out back through the hole, then mend the hole with a pellet of clay -- operations which are quite different from her normal building routines. Hingston [30] has described in detail the actions of another wasp -- Rhynchium nitiderium -- in repairing a man-made hole in a clay-pot. On one occasion the female tried for two hours to mend the hole with bits of material taken from the wall of the pot. Then night came and she had to give up. Next morning she flew straight to the damaged spot and set about repairing it by a different strategy. In the normal course of events the wasp works from outside. But now, in order to repair the hole, 'she examines it from both sides and then, having made a choice, elects to do the repair from within'. [31]

  Equally surprising is the ingenuity, of the caddis-fly larva. If a group of larvae are ejected from the tubular 'houses' which they built, and are then allowed to return, they often get mixed up and enter the wrong 'house' which is either too big or too small. The larva then sets about to cut off parts of the tube or to add to it, until it fits it exactly. Again the 'consummatory acts' in these activities are quite different from those in normal building.

  Many birds, too, are capable of such 'super-flexible' behaviour in emergencies. If their brood is taken away, they will re-start their sexual cycle, court and mate out of season, and get a new family going. In some species, in the absence of the female, the male bird takes over her duties in feeding the young -- which never happens under normal conditions.

  Lastly, a brief mention must be made of 'supra-individual codes' -- such as those which regulate activities in the honey-bee hive. Lindauer, among others, has shown that 'the programme of work carried out by the individual is not determined by the physiological state of the insect but is dictated by the needs of the colony as a whole'. [32] An individual worker-bee hardly ever builds a complete cell. She may start the cell, with wax from her own glands, then complete another cell, started by a colleague, using her own wax or that of another bee -- whichever happens to be convenient. Generally, there is rigid division of labour according to age groups: each worker has to perform a different kind of 'National Service' in different periods of its life. During her first three days, she works as a cell-cleaner. For the next three days she feeds the older larvae with honey and pollen from the stores. Then she feeds the younger larvae, who get an additional diet -- a liquid secreted from glands on the worker's head. At the age of ten days she is engaged in complex household chores and building activities. At twenty days she takes over guard duties at the entrance of the hive; and finally, she becomes a forager and remains one until the end of her life. But even among the foragers there is further specialization of labour: some of them become 'scouts', whose task it is to discover new sources of food, and to communicate, on their return, the nature and location of it, in their dance-language, to the hive. [33]

  But this is not all. If one of the specialized age-groups is artificially eliminated from the colony, a kind of collective super-flexibility manifests itself in the hive: other age-groups deputize for the vanished group 'and thus save the superorganism. When, for instance, all pollen-and-honey foragers are taken away -- usually bees of twenty days or over -- young bees of scarcely six days old, who normally feed the larvae, fly out and become foragers. If all building workers are taken away -- those between eighteen and twenty days old -- their task is taken on by older bees who had already been builders before, but who had gone on to the stage of forager. To this end they not only change their behaviour, but also regenerate the wax-glands. The mechanisms of these regulations are not known.' [34]

  Thus at one end of the scale we find rituals, fixed action-patterns, vacuum and displacement activities -- rigid, automatized, and compulsive, petrified habits of unknown phylogenetic origin. At the other extreme we find supra-individual codes which govern behaviour of remarkable flexibility, and original adaptations which lie outside the animal's normal skills and habit repertory. In all forms of social organization -- from courtship, mating, and fighting rituals, through territorial demarcations, up to the complex insect state, we find an interlocking of individual behaviour-patterns into a collective super-code which casts the individual bird or bee into the role of a part in the social whole. Thus we see the hierarchic part-whole relationship repeated on the level of social organization, where the integrative functions of catalyzers, inductors, and nerve impulses are superseded by interlacing systems of social releasers, including communication by signs and symbol -- from display, through bird-song, to the dance-language of the honey-bee.

  NOTES

  To p. 481. The equivalent of the term 'appetitive behaviour' in American behaviourist theory are Hull's drive-stimulus (Sd); and his 'fractional antedating goal-stimuli and responses' (SgRg).

  To p. 481. Out of this grew the theory that the fixed pattern of the consummatory act -- and not the 'appeted stimulus' -- is the goal of the animal's striving and the source of the 'action-specific energy' of the drive; but the subject is outside the scope of this book.

  VII

  IMPRINTING AND IMITATION

  So far we have discussed the codes of morphogenesis and innate behaviour, which emerge ready-made from the black boxes of evolution' -- like All Baba's thieves, popping out of the urns in which they were hiding.

  In the chapters which follow we shall discuss the ontogenesis of behavioural codes -- the acquisition of habits, knowledge, skills, by the processes' of learning from experience.

  The 'Following-res
ponse'

  The transition from innate to learnt behaviour is sharply highlighted in the phenomena of imprinting. The follow-the-leader response of the gosling is governed by an innate code; it must stick to the mother-goose or perish. But like many phylogenetically acquired codes it seems to have been formed according to the principle of parsimony. It can be triggered off by any releaser which satisfies very broad Gestalt criteria of 'goose-likeness' -- including German ethologists and even inanimate moving objects of a certain size. In a normal environment this would indeed be sufficient to ensure the gosling's survival, since the first sizeable moving creature seen would be the mother-goose. Accordingly, a young goose, reared from the egg in isolation (or in the incubator) will accept -- during the brief critical period of maturation when imprinting occurs -- its human keeper as its 'mother', and follow him around. Once this has happened the process becomes more or less irreversible: the 'imprinted' bird will reject the company of other geese and attach itself only to memebers of the human species -- treating them as parents, companions, and later on as objects of sexual advances. Many other birds, and possibly also some fish and insects, show the phenomena of imprinting in varying degrees.

  Here, then, we have a pregnant example of the genesis of a matrix through the integration of innate and acquired behaviour-patterns. The built-in 'following response' has the characteristic autonomy of motor-patterns which we have met before: it is triggered off and modified, but not created by the environmental input. The first step in the development of the matrix is the act of imprinting itself; it must occur, as already mentioned, during the critical phase when the young bird is susceptible for it (in ducks, for instance, between eleven and eighteen hours after birth, with a pointed peak in the susceptibility curve at sixteen hours). [1] The input which triggers off the following-response is at this stage an undifferentiated and primitive sign-releaser: 'Large moving object' -- much simpler in character than the more specific Gestalt stimuli which release the fighting or mating instinct in the stickleback ('red belly', 'swollen belly') or the begging response of the herring-gull chick ('red spot on beak').

  The next stage is one of perceptual learning. After a few hours, even a few minutes, of following a human being, the gosling will follow only human beings -- it has somehow learned to 'abstract', or 'encode' in its memory some specific Gestalt-characteristics of homo saplens which distinguish it from other 'shapes that move'. On the other hand, at this stage all human beings are still 'equipotential' members of the emergent perceptual matrix. At a still later stage, the goose may become attached to one or more single individuals, that is to say, it learns to discriminate individuals within the species -- as, vice versa, animal breeders learn to sharpen their perception and to distinguish one sheep or goose from another.

  We thus meet, already on this level, the twin phenomena involved in all learning processes: generalization ('transfer', 'abstraction') and discrimination (segregation of pattern, selective inhibition of responses to non-specific stimuli). These basic processes will be discussed later (Chapter X); in the meantime, let us note that the innate, primitive 'rule of the game' which made the new-born animal respond to 'things-that-move', has been sharpened and elaborated into a more complex set of rules by a series of 'steps. Each of these steps involved a restructuring of the perceptual matrix by successive generalizations and discriminations -- which we may regard as quasi-extensions of functional integration and structural differentiation into the learning process. Morphogenesis and learning form continuous series which overlap during maturation; and the matrices of innate and acquired behaviour form an equally continuous hierarchy.

  Bird-song and Parrot-talk

  There are no sharply defined boundaries between imprinting and learning by imitation, or by trial and error. The word 'imprinting' itself is a translation of Heinroth's Prägung, [2] by which he meant to indicate the dramatic form of learning in birds which we have just discussed. Its chief characteristics are: it is species-specific and directly dependent on innate organization; quasi-instantaneous; and limited to a relatively brief period in the animal's life. By applying these criteria, Thorpe has extended the concept of imprinting to include 'a bird's instantaneous attachment to territory, its occasional attachment not only to humans but also to other animals and even inanimate objects; and lastly, the song-bird's way of learning its species-characteristic song'. [3]

  Apparently in some birds such as thrushes, warblers, pippits, the whole song is genetically 'built in' and can be but slightly modified by learning; while in others, for instance the skylark, it is mainly learned. In chaffinches Thorpe has shown that 'while the basic pattern of the song is innate, all the finer detail and much of the pitch and rhythm have to be acquired by learning.'* We have here another example of a 'roughed-in' pattern (p. 470) whose details are filled in later by that particular type of 'feedback' process which constitutes learning.

  When we turn to imitative bird-song and parrot-talk, the part played by innate organization is obviously less specific and the part played by learning much greater; yet the difference is again one of degree. In fact, Prägung means stamping (a coin), which makes the continuity between imprinting and 'stamped-in learning' even more obvious.** Less obvious, however, is the biological purpose or adaptive value of the striking capacity of parrots, mocking-birds, starlings, etc., to imitate the songs of entirely alien species -- including 'God Save the King'. [4] Now parrots living in freedom in their natural environment utter only a few fixed, simple types of cries; yet folklore apart, we have no lesser authority than Lashley describing a captive parrot with a 'vocabulary' of between fifty and a hundred words; and there is reliable evidences that both parrots and robins can learn to utter certain words meaningfully. Granted that vocal imitation, as McDougall has pointed out [6], is a special case owing to the close integration of auditory-vocal patterns, one must nevertheless admit that such imitative ability is 'a further example of pre-adaptation for apparently remote and unlikely contingencies, specialization going in advance of immediate adaptive requirement, and as such on a par with the astonishing number-sense which can be developed in many species by careful training. Such a counting ability seems to offer even less practical advantage for a wild bird than the features we have been considering; all are as yet somewhat mysterious.' [7] (The counting ability of birds was revealed in Otto Koehler's famous experiments, to be discussed later.) Ethologists such as Koehler (not to be confused with Wolfgang Köhler), Lorenz, Craig, and Thorpe all agree that the tonal purity, the 'inventiveness' and improvisation in the advanced forms of bird-song should be regarded as 'the first steps in both music and speech'. [8]

  To mention one example among many: Waite, at the Museum in Sydney, owned an Australian magpie whom he taught by playing on the flute 'a fifteen-note melody in two distinct phrases'. Some years later he got a second magpie which learned the tune from the first. The two birds then developed the habit of singing it antiphonally, the first singing the first phrase, and the second only the second. 'Later the second, younger, bird died whereupon the first resumed its performance of the whole'. [9]

  Examples like this show not only the great flexibility of these auditory-vocal matrices. They also show that the total pattern -- the rudimentary code of the hand-reared chaffinch -- develops first, and that learning the song does not consist in the chaining of individual notes according to the S.-R. scheme, but in the elaboration and variation of the pattern. We further note that originality or 'inventiveness' make their appearance at only a few removes from innate and imprinted behaviour. Lastly, the striking learning abilities of some birds, which are only revealed under the abnormal conditions of captivity -- these examples of 'pre-adaptations for remote and unlikely contingencies' remind us of regenerative potentials manifested in response to traumatic challenges.

  Untapped Resources

  We have seen evidence of this latent super-flexibility -- of 'doing wonders' in adversity -- on every level: from the restoration of locomotive patterns i
n mutiliated insects and rats, through the emergency redistribution oflabour in the beehive, up to the solution of blocked problems by 'thinking aside'. In recent years, unsuspected learning abilities were revealed in such widely different classes as fiatworms, dolphins, and seals -- the latter, apparently, can even be taught to obey visual sign-commands printed on cards. Yet the evidence for a surplus, or reserve, of learning potentials far surpassing immediate adaptive needs has always been there in our own species: ten thousand years ago our ancestors fought with clubs and arrows, but the structure of their brains was the same as ours, and therefore potentially just as capable of learning Boolean Logics or the principles of making a nuclear bomb. Even the dumb fish have been shown to have optical capacities for form and colour discrimination far in advance of their needs under natural conditions. Thorpe comments: 'It does indeed seem to be a general feature of animal life that the precision and sensitivity of sense organs is higher than the environment would appear to justify. This fact poses a serious problem for students of evolution, since it is not easy to account for such perfection on the basis of natural selection alone'.*

  It seems that this overshooting of the mark, this giving more than was asked for, is an inherent characteristic of the mechanism of evolution. In homo sapiens the 'overshooting' is demonstrated by the fact that mental evolution -- learning to exploit the surplus potentials in his brain -- has been going on for an astronomical period, and with no end in sight. The problem is not so much why mental evolution occurs in man, but why no similar phenomenon -- learning to use their native equipment to maximum capacity -- seems to have occurred in any other species, although many animals demonstrate the existence of their untapped resources in captivity. (Animals, it is true, keep no written records of their discoveries, but these could have been transmitted by imitative learning.) All this is a nice subject for speculations on a rainy day; the important point in our context is the hard core of evidence to show that various animals reveal in captivity various degrees of originality and resourcefulness which are not displayed under natural conditions. Most animals seem to have more sensitive organs than they need, and more latent capabilities than they ever learn to actualize, except when challenged under propitious circumstances. That bar-pressing experiments with rats are not the type of challenge designed to elicit original responses, need not be emphasized, and it is not surprising, therefore, that leading Behaviourists have either denied the occurrence of original responses or put them down to chance. We shall return to the subject in later chapters.