INSTINCT, INTELLIGENCE AND APPETITE. BY A. CAMPBELL GARNETT. (From the Psychological Department of King's College, London.) I. Instinctive behaviour as always characterized by intelligence (p. 249). II. The function of relations in instinct process: the theory of 'configurations' or Gestalt (p. 252). III. The distinction of appetitive from non-appetitive instinct processes (p. 259). IV. The distinction between instinctive tendencies and general mental energy (p. 261). I. INSTINCTIVE BEHAVIOUR AS ALWAYS CHARACTERIZED BY INTELLIGENCE. It has been customary with perhaps the majority of psychologists and philosophers to draw a distinction between instinctive and intelligent behaviour. It is the principal purpose of this paper to attempt to show that this distinction is wrong and to point out what appears to be the true relation between instinct and intelligence. The distinction goes back, in modern psychology, to the time when human actions were believed to be all guided by reason in the light of prospective pleasure and pain, and, perhaps, of the 'moral sense,' while the behaviour of animals was directed automatically by instincts which chance or the Creator had given them. This view of animal instinct as being perfect, rigid, unintelligent and automatic has, however, been as widely shaken by modern investigation as has the complementary belief in man's sublime rationality. We are forced to make distinctions even among those animal activities which seem to be innately conditioned and to recognize the existence of similar innate tendencies to action in man. Some of these innately conditioned actions are mere responses of a part of the organism to some specific stimulus and are probably as rigidly conditioned and unintelligent as all animal instinct was once thought to be. Others are responses of the whole organism to a specific stimulus and perhaps some of these (tropisms) may be completely unintelligent, that is, involve no cognitive processes. But we do not have to rise very high in the scale of animal life before we come to creatures whose responses to certain aspects of their environment, while undoubtedly innately conditioned, show indubitable marks of intelligence; i.e. cogniMed. Psych. VIII 17 tion plays a part both in the reception of the stimulus and in the subsequen process of response. I propose to limit the term 'instinct' to these cases and to define instinctive behaviour briefly as "an innately determined response to the cognition of a meaningful situation." This working definition will hold good even though there are many borderline cases where we cannot decide as to the presence of cognition and even though we do not know what cognition is. The definition would still be valid even if cognition were nothing more than complex molecular change in the cerebrum, as the Behaviourists say it is; though the theory which I shall proceed to present as to the functioning of cognition in instinctive behaviour is, I think, not consistent with such a view. If cognition plays a part in all instinctive behaviour this means that the older distinction between instinct and intelligence was wrongly drawn. Instinctive behaviour is always more or less intelligent. Certainly, in a creature of low intelligence, any type of instinctive activity (a) is greatly limited in the kind of stimulus that can evoke it, and (b) is similarly limited as to the kinds of behaviour with which the creature can respond. But cognition is always present in some degree in every response or activity to which the term 'instinctive' can rightly be applied. When such behaviour becomes very intelligent and sophisticated, we are, however, apt to lose sight of the fact that it is instinctive. But the man who writes a story to earn money to buy food is not, strictly speaking, acting less instinctively than the ox that crops the grass; he is acting more intelligently; that is all. It is not the details of the response that are fixed by the innate factor we have called instinct, but rather, as we shall see, the general nature of the end towards which the response shall move; the details are fixed by the limitations of the creature's intelligence and the structure of its sensory-motor mechanism. But since both the stimulus and the mode of response in instinctive activity are so greatly varied and modifiable as is suggested by this comparison between man and ox, how, it may be asked, are we to say of any such piece of behaviour that it is a manifestation of such and such an instinct? How distinguish instincts from each other? With this problem I have dealt on a previous occasion1. May it suffice to say here that I think the most important, indeed the one really satisfactory criterion, is that of the nature of the end-experience in which the conation finally ceases, and satisfaction is found. Early attempts to distinguish the instincts by reference to stimulus and response have led only to confusion in the creation of a multiplicity of alleged instincts, 1 British Journal of Psychology, Jan. 1928, while McDougall's suggestion1, that the emotion marking the instinctive process may be taken as criterion of the instinct to which it belongs, is of value only with regard to a certain limited number of instincts. With some of the most important instincts its quality is too vague for a distinguishing mark. There is only one uniform characteristic of every distinct type of instinct process which is different for every such different type. That is the end-experience in which the creature finally finds satisfaction. This end can be known in introspection and safely inferred in observation of animals. By reference to it we can distinguish the kind of instinctive drive in operation even in species very different from ourselves and in our own species even when the actual activity is of a highly sophisticated nature. This end-experience, like the stimulus, comprises the cognition of a meaningful situation. I use the term 'meaning' here in the sense of 'overknowledge,' that which is known, in a percept or concept, over and above the actual lived experiences which are known. In this sense the 'meaningful situation' includes that which is known (1) of the relations subsisting between the objects cognized, (2) of correlates and other associations to those objects, (3) of relations between these, (4) of the relation between the knower and the object known. In man this 'meaning' may be very full; in the animals it is probably always very scanty. An instinct process, then, begins with the cognition of one meaningful situation and ends with the cognition of another. Viewed introspectively it contains three phases: (a) the stimulus, which consists in the meaningful situation cognized; (b) the resulting active or transitional phase, which is experienced as one of a certain mental tension, or excitement, during which attention is held by certain objects, or a certain succession of objects, which the individual is impelled to deal with in a certain way more or less clearly seen to promise satisfaction; (c) the end-experience of satisfaction, relaxing of tension and cessation of conation. Observation of instinctive activity in animals gives evidence of the presence of the same three mental phenomena of cognition, attentive striving, and satisfaction and relaxation. There is reason therefore to believe that the introspective evidence gives us the true view of an instinct process, though allowance must be made for vast differences in the level of intelligence displayed by different species as affecting in particular the amount of foresight which enters into the process and the experience2. 1 W. McDougall, An Introduction to Social Psychology. 2 F. Aveling, article in the British Journal of Psychology, April 1926. II. THE FUNCTION OF RELATIONS IN INSTINCT PROCESS: THE THEORY OF 'CONFIGURATIONS' OR GESTALT. In considering the function of intelligence, i.e. of cognitive processes, in instinctive action, it is of the first importance to remember that the stimulus is never merely a sensation or congeries of sensations but always a relatively complex situation. A species of spider which has an instinctive tendency to flee from a bee has been shown, says Koffka', not to be stimulated to this reaction by any particular colour, odour, or size, but yet to take to flight when the real bee is presented even in the most unusual positions. Now a bee from certain viewpoints presents a very different visual stimulus from that in others. The effective stimulus to the spider then would appear to be certain sensory stimuli in a certain system of relations. Birds that are accustomed to the sight of human beings and have occasionally been shot at will often not take to flight at sight of a man with a stick beneath a tree, but will do so immediately if he should point the stick at them. Similarly birds in a game reserve will allow human beings to approach very close, but the same birds, when they are found outside the reserve, will not allow a man to approach anywhere near gunshot range. These latter are examples of acquired instinct stimuli but are none the less important for that. They show decisively that the relations which enter into the perception of the object are of crucial importance to its effect on the creature. An alteration in those relations is an alteration in the meaning of the situation. process. This does not mean, of course, that these creatures educe the relations existing between the lived experiences which initiate the instinct Perhaps only human intelligence is capable of actually educing relations, which involves attending to relations as such. What these facts do seem to imply, however, is that what the animal cognizes is a system of related experiences, and that it is the fact of the relations subsisting between the elements of experience, rather than that of the visual, auditory or olfactory qualities which characterize them, that constitutes the cognition of them the beginning of an instinct process. The lived experiences as cognized form a certain 'configuration,' or, as the Germans say, a Gestalt, and it is this that constitutes them an instinct stimulus. This very fruitful conception of 'configurations' has been submitted to searching criticism by a number of writers. In opposition to the Gestalt School, Rignano, in some recent articles2, maintains, I think 1 Koffka, The Growth of the Mind, rightly, what he calls "the qualitative autonomy of the sensorial elements," and contends that it is the affective reaction of the individual to these sensorial elements that gives them their unity and distinctness as separate 'things' in cognition and brings about the appropriate correlated behaviour. But the notion of 'configuration' is too valuable to be easily dispensed with, for such facts as those referred to above in reference to spiders and birds show clearly that the relations of the various sensorial elements to each other do play a part of tremendous importance in even these very simple acts of cognition. Yet the meaning of a situation cannot consist merely of the cognized spatio-temporal relations subsisting between the sensorial elements. A barking dog and a silent man with a gun have very similar meanings for a rabbit, yet one cannot possibly see how those two sets of sensorial elements could possibly be fitted into the same set of spatio-temporal relationships. Neither can the meaning be more easily explained as due to affective reactions to the qualitatively autonomous sensorial elements, for the sensorial elements are so vastly different that the number and selectiveness of the paths of association would have to be incredibly enormous if the cognition were effected in this way. For it must be remembered that, on Rignano's view, it is the affective reaction to the sensory stimuli that serves to separate and unite those sensorial elements into meaningful things. So the affect must be aroused before the thing is cognized as such; i.e. the appropriate affect must be aroused by the mere mass of stimuli. But these stimuli may be so enormously varied, and such very similar sensory stimuli may produce such very different results when cognized in slightly different relations, that the hypothesis would require an incredible multiplicity and selectiveness of the association paths. It is the great merit of the Gestalt psychology to have shown that the relations which enter into the cognition act as an intermediary. Certain sensory stimuli, having certain relations to each other, tend to be cognized in those relations, i.e. to stand out in consciousness as a related system, a 'thing.' It is these cognized systems of relations that, on the Gestalt view, are correlated with behaviour, and this reduces the number of association paths required to a quite reasonable number. But there still remain the problems (a) of finding a place in the Gestalt theory for the apparent effect of the affective-cognitive factor in the determination of meaning; (b) of discovering some element of similarity in the Gestalt arising from two very different sets of stimuli which nevertheless stimulate the same instinct. I would suggest that these two difficulties in the Gestalt theory may |