The Animal Mind: A Textbook of Comparative Psychology

was found by Hesse to give no further response to sudden shadows when the stimulus was frequently repeated (321). Hargitt (285) reports the same of tube-dwelling annelids. Von Uexkiill reports that the sea-urchin Centrostephanus longispinus ceased to respond to shadows after three successive stimulations (736). Nagel observed that certain eyeless mollusks which react to sudden darkening very quickly get used to the stimulus and cease to respond; often after one reaction they decline to react for several hours.!. The mollusks that responded to sudden bright- ‘ening rather than to shadows, that were in Nagel’s phrase | photoptic rather than skioptic, took longer to become ac- | customed to repeated stimulation, but did so by gradually weakening their reaction (520). A web-making spider ‘that was found by the Peckhams to drop from its web at the sound of a large tuning fork declined to disturb itself after the stimulus had been repeated from five to seven times (570). Ants “become used” to the ultra-violet rays which they ordinarily avoid (220). The responses of dragon fly nymphs to light are less marked as the stimulus is repeated (636), and the same is true of mosquito larve

Where such an effect as this is temporary, the most obviously suggested cause for it is fatigue. In our own experience this word is used chiefly with reference to motor processes; we perceive a certain signal, but are too fatigued to respond. On the sensory side, when a repeated or con- Lismued stimulus is no longer perceived, we call the phenom- 1 The opposite phenomenon is reported by Rawitz of the mollusk Pecten, whose response to a shadow was the shutting of its shell. Repeated or long- ' continued shadowing, instead of deing away with the reaction, caused the animal to remain with closed shell for a long time; an intensification of the

reaction which suggests the effect of summation of stimuli (628). We may infer that the stimulus in such a case is injurious. enon one of adaptation. In true sensory adaptation, the sense organ becomes incapable of responding to the stimulus; for example, a person who has been for some time subjected to a certain odor is unable to smell it any more, however much he tries. Closely related to this phenomenon and yet different from it, is the lapse of attention to a; repeated stimulus: we no longer notice the ticking of a clock, although the sense organ is unaffected by its con-

tinuance, and we can quite well hear it if our attention is\ attracted in that direction. That the failure of Stentor to respond to successive stimuli is not due to motor fatigue appears quite certain to Jennings, since under favorable conditions he has obtained reactions from the animal for a period far longer than that occupied by the process of getting used to slight mechanical stimulation (370). And in most of the cases cited, the acclimatizing process seems to occur too rapidly to make fatigue of the motor apparatus probable. In the lowe animal forms, sensory adaptation offers the most natural explanation for the phenomenon; in the higher animals lapse of attention is very likely also involved. The modi, fication of consciousness in both cases would be the loss’ of the sensation; where adaptation occurs, the sensation’ would be for the time irrecoverably lost; where there is merely lapse of attention, it could be regained by a proper direction of attention.

A much discussed case of the cessation of response to a repeated stimulus is found in connection with the foodtaking reaction. One would expect the dominant condition here to be loss of hunger, and as a matter of fact, observers of the feeding processes in many lower animals have found that such reactions cease or turn into negative responses when the animal is satiated; although Piéron indeed reports that while the responses of Actinia equina and A. rubra to mechanical stimulation cease on repetition of the stimulus, those to food stimulation continue indefinitely (581). If the change from food-taking to negative reaction has a conscious accompaniment, this might naturally be thought of as a change from pleasant to unpleasant affective tone. Nagel observed that if a ball of filter paper soaked in fish juice were placed upon one of the tentacles of the sea-anemone Adamsia, it was seized as eagerly as a ball of fish meat, but that when this deception had been several times repeated, the ball was held for a shorter period each time, and was finally rejected as soon as offered. Nagel is inclined to think that this is learning by experience, and points out that the psychic life of Adamsia must possess little unity, for the ‘‘experience” of one tentacle does not lead other tentacles to reject the paper balls at nce (521). Parker finds similar behavior in Metridium, and explains it by saying that the filter paper offers but a weak food stimulus, and that ‘the successive application of a very weak stimulus is accompanied by . . . a gradual decline in the effects, till finally the response fails entirely” ; in other words, that we have adaptation to a food stimulus (533). Jennings fed Aiptasia alternately with pieces of ;crab meat and with filter paper soaked in meat juice, the | result being that the fifth piece of filter paper was rejected ,-—— but so was the crab meat thereafter. Jennings came to the conclusion that the phenomenon is due simply to loss of hunger on the animal’s part, and that where Parker ound that the crab meat would be taken after the filter per was refused, it was because the latter was a weaker stimulus and naturally was the first to call forth the effects of satiety.

The objection to the hunger hypothesis is that other tentacles of the same animal will react after one tentacle has stopped; satiety ought surely to affect the entire organism (374). Allabach, in the light of these researches, made a careful study of Metridium. She disposes of the psychic learning by experience theory of Nagel by saying that the only experience upon which the animal could reject the filter paper must be experience that it is not good for food. This could be learned only by swallowing it; but the failure of the reaction occurs just as well when the animal is prevented from swallowing the filter paper. That the phenomenon is not one of adapta~ tion to weak stimuli is shown by the fact that it may be brought about by successive feedings with meat which is not allowed to be swallowed. It cannot be due to loss of hunger, for this is experimentally shown to affect all the tentacles at once. Allabach concludes that it is simply a case of local fatigue of the tentacles. The taking of food by a tentacle involves the production of a considerable}! quantity of mucus, the immediate supply of which is prob/ ably exhausted after a few reactions, and a short period of rest is required (3). Parker (551) is still of the opinion =

Another case of the cessation of reaction to a repeate stimulus is reported by Wasmann of ants in an artificial ‘nest, which assumed the fighting attitude in response to the movement of a finger outside the nest, but after two or three repetitions of the motion were no longer disturbed (762). Where animals as high in the scale as the ant and’ spider are concerned, it is possible that this process of getting used to a stimulus may involve rather a dulling of emotion than a disappearance of sensation. This phenomenon also is familiar in our experience, and may be called emotional adaptation.

(is caution As Jennings suggests, if the sea-anemone that contracts at the first ray of light were to remain contracted in steady illumination, it would lose all chance of getting food under the new conditions (374). The negative reactions ordinarily involve interruption of the food-taking process, and it is important that they should not be continued in response to stimulation that is relatively permanent. Hargitt thinks that the loss of reaction to repeated shadows which he observed in marine worms may be an adaptation to the varying illumination caused by ripples at the surface of the water (285).

A very important psychological question concerns the permanence of the effects of adaptation. Sensory adaptation and lapse of attention to repeated or continuous stimuli, as these phenomena are met in our own experience, are not considered phenomena of learning at all. The former is purely temporary in its effects: the person who has become so used to an odor that he cannot smell it shows o effects of this experience half an hour later. The effect of familiarity on emotion and on attention is more lasting : one’s loss of attention to a clock ticking in one’s room may persist despite more or less prolonged absences from the room, although a sufficiently long absence, during which one encountered no ticking clocks, would cause the sound to be noticed again. The loss of emotional response to a familiar stimulus may persist for some time. Emotional adaptation and lapse of attention to continued stimuli jmay fairly be termed learning in proportion as their effects re more than temporary.

actions last over a considerable interval between the stimuli. This seems to be increasingly the case, the higher the animal. Thus Hydra, which is only a ccelenterate, if it is allowed to reach full expansion after having contracted at a touch, will respond to the second touch just as it did to the first; the stimuli, to exert any influence on later reactions, must come in quick succession. On the other hand, in the responses o mollusks to shadows, the experiences of one day appear to extend their effects to the following day (520; 588, 590). |' Here we are dealing with a new type of modification by experience, though one which develops directly out of sensory adaptation ; namely, the relatively permanent dropping off of useless movements.

§ 73. Modifications Due to Relatively Permanent Effects of Stimuli In true learning, the conscious experience and the be havior of an animal suffer changes so lasting, relatively speaking, that they cannot be set down as due merely to adaptation of the sense organ, muscular fatigue, hunger, satiety, or any other variable physiological state of the organism. On the other hand, as we saw in Chapter IT) the modifications must occur rapidly enough so that there is not time for actual changes in the animal’s muscular structure to be produced. In animals which possess nerv ous systems, true learning is probably always the result of alterations in the connections between the elements of that system, such that the nervous process is able to pass easily in a direction where it originally encountered high resistances.

The fundamental law of all learning is the Law of Repeti- \ tion, whereby when a nervous process traverses a certain | pathway in the nervous system, it leaves the resistances in — that pathway less than it found them. This is the law in accordance with which, when we wish to learn anything, we repeat it over and over, relying on the certainty that each repetition will make the next one easier. With this law in mind as an essential postulate, we shall survey the types of true learning found in the lower animals under the ollowing four heads: (1) learning involving the dropping out of movements; (2) learning involving the formation of series of movements; (3) the recognition of landmarks ; (4) learning involving the anticipation of movements.

Among all the movements which an animal is capable of making, there are some which are closely connected with the great needs of its existence, and others whose connection with such needs is only indirect and casual. The general process of adjustment to environment which has made _ the animal what he is, has so ordered matters that the '| vitally important movements are in a state of especial ‘readiness to be performed. The nervous resistances along the pathways leading to the muscles used in these movements are congenitally low. Such responses are what ‘Sherrington (681, p. 229) has called ‘‘prepotent reflexes.” Now if we survey all the cases in which an animal learns

, by experience, we are obliged to conclude that on some ' principle of economy of energy, zsolated movements which do not bring any consequences of importance to the organism nd to be dropped, and their places taken by a state of rest. his seems to be the law according to which we ourselves cease to pay any attention to our familiar surroundings. e cease to notice the ticking of a clock, although no adaptation takes place in the ear itself; we sleep undisturbed by the noise of the trolley cars which is distracting to our friends from the country. The spider experimented

on by the Peckhams reacted each day to the sound of a tuning fork by dropping from its web until the sound had been repeated some half dozen times, but after the fifteenth day it would not drop at all (570). Piéron (588, 590) found that snails, while at first responding to shadows by withdrawing the tentacles, on successive days stopped reacting after fewer and fewer trials; and believed he could trace a parallel between the laws of this learning and those of human memory. There is no question in such cases of the reaction’s being dropped off in favor of some other-reaction. It is dropped off, as it were, by its own weight; simply because it is useless. This same principle seems to enter as a codperating factor in cases where animals acquire a discrimination between stimuli. The apparent ability of sea-anemones to distinguish between real food and filter paper soaked in foodjuice (see page 254) is, as we have seen, ascribed by some | to sensory adaptation, but the experiments of Fleure and Walton (228), if their results are accepted, would indicate | that true learning is involved. They tested Actinia with a scrap of filter paper once every twenty-four hours, placing it on the same tentacles, which usually carried it to the mouth, where it was swallowed and later ejected.

swallow the fragment, and in two more days the tentacles refused to take hold of it. Other tentacles could be “deceived” at least once or twice after this, but very soon manifested the inhibition. All traces of the learning weré lost after from six to ten days interval. Another anemone, Tealia, learned more quickly than Actinia. Again, Herrick (297) found that catfish, when the barbels were touched ‘with a bit of meat, immediately seized it. If a piece of cotton wool were used instead of the meat, they made the

same reaction, but after this experience had been repeated a certain number of times they ceased to respond to the cotton, although they still took meat eagerly. The point which especially concerns us is this: ‘I rarely,” says Herrick, ‘‘after the first trials, got a prompt gustatory reflex with the cotton.” The learning persisted for a day or two. The axolotl learned in a similar way to discriminate between pieces of meat and pieces of wood (276). Hermit crabs, which when young try to take up their abode in all sorts of unsuitable objects, glass balls, for instance, later in life make no such efforts (194).

Whether or not a movement which brings no favorable results will be dropped off and a state of no movement will take its place depends on how strongly prepotent the moveent is; upon the strength, that is, of the innate tendency to riter on color discrimination in the creek chub, our first ( method failed because it required the dropping off, as use- \less, of a strongly prepotent reaction, and the substitution \of no response at all. Red forceps and green forceps, each containing food, were plunged one at a time into the water ; the fish was allowed to get the food from the red forceps, but the green ones were withdrawn before it had a chance to bite. The time which the fish took to rise and snap at the forceps was measured by a stop-watch, and in the course of 131 experiments the fish had not learned to rise to the green any less promptly than to the red. In other words, no tendency to drop off the useless movement of rising to the green was detected, although later experiments Showed that the fish could distinguish between the two forceps. The movement of rising to and biting any small ‘\object in the water was so vitally important to the fish ‘\that it could not be dropped off (757). On the other

hand, Thorndike (708) successfully carried out this kind of training with Cebus monkeys: both of his subjects learned’ to come down to the bottom of the cage for food when the experimenter took the food in his right hand, and to stay up when he took it in his left hand, the food being withheld if the monkeys came for it in the second case. Cole (134) trained raccoons to climb up on a box for food when one of » two differently colored cards was shown, and to stay down | when the other one appeared, by not feeding the raccoons if they climbed up for the wrong card.

The dropping off of movements takes place with morel speed and certainty if they are made to give place, not simply to a state of no movement at all, but to a movement of greater prepotency than their own. Especially effective in\ thus causing the elimination of a movement is the negative} reaction of withdrawal from injury. Thus if a movement A results in actual harm to the organism, the harmful stimulus thus produced brings about the negative response; and the negative reaction is as a rule prepotent over all others, The next time the movement A is initiated, the negative reaction is also initiated, and being prepotent, it is abl to check effectively the performance of movement A. Thus we have the dropping off of harmful movements, a process which is in evidence whenever punishment is used in studying the learning power of animals. It also appears when a successful negative reaction permanently takes the place of unsuccessful ones. We saw in the first part of this chapter that when an animal is repeatedly subjected to a strong and harmful stimulus, it goes through a series of reactions, all directed to getting rid of the stimulus, until one is finally successful. Now if this process is shortened in successive trials, so that the successful negative reaction comes to be made at once and the unsuccess-,

ful ones dropped off, we have a case where the dropping off is not simply of useless but of harmful movements (since the unsuccessful ones all result in a repetition of the harmul stimulus); the final state is not one of no movement, \but of victory for the successful negative response. A very interesting illustration of this type of learning was obtained from Paramecium by Stevenson Smith (688) and by Day and Bentley (178). The method used by these experimenters was fundamentally the same. A glass tube was drawn out until it was so fine that not more than one Paramecium could get through it. This tube was filled with water up to a certain point, and a single Paramecium, carefully isolated for identification throughout the experiment, was allowed to swim up the tube until the surface film was reached. The animal behaved towards the film as to any mechanical stimulus, darting backward, rolling over towards the side away from the mouth and swimming forward again. Since the tube was so narrow, this method, which ordinarily succeeds in avoiding obstacles, brought the animal against the surface film again. After repeatedly going through the same performance, the Paramecium varied its response and succeeded in turning completely around in the tube by bending its body double. On being put again into the same predicament, it gradually diminished the number, of trials of the unsuccessful negative response, and arrived at the point where it almost immediately doubled over on striking the surface film. These observations established the existence | of a relatively high type of learning in the simplest group

\of animals. ' In this case the movements that are dropped off are \themselves negative reactions. In other cases they may e feeding reactions or other responses whose vital importance, though great, cannot compete against that of the negative response called forth by their injurious effect in the special case. Learning by punishment is in most caseg\) especially rapid. Its effect may be to inhibit altogether, for some time, a certain instinct. For example, the ex perience of receiving an electric shock when they seized a certain kind of food prevented frogs from feeding at all for several days (657). Rats which were being trained to discriminate between a lighter and a darker passage with the use of an electric shock acquired a distaste for the apparatus as a whole (328). Mébius in 1873 (497) made some experiments with a pike, afterwards repeated by Triplett (720) with perch, which illustrate the same phenomenon. ‘The fish was kept in one half of an aquarium, separated by a glass screen from the other half, in which minnows were swimming about. The pike naturally dashed at them, and whenever it did so bumped its nose on the glass partition. After a considerable period of this sort of experience, the glass screen was removed, and the minnows were allowed to swim freely around the pike, when it was found that the latter’s instinct to seize them had been wholly suppressed by the harmful consequences of such action. Here, again, the chances that a movement will be suppressed in favor of the negative response depends on how great the degree of its prepotency is. It was a rash conclusion on Bethe’s (49) part to deny the learning ability of the crab because, although every time it went into the darkest corner of its aquarium it was seized by a cephalopod lurking there, it did not in six such experiences , learn to inhibit its innate tendency to avoid light: further training would probably have been successful. Yerkes (822) trained an earthworm, by giving it an electric shock when it followed its innate inclination for turning towards

a darkened region, to turn away and towards the light. The cockroach, as is well known, prefers darkness to light: Szymanski (7o1), however, succeeded, by giving it an electric shock when it ran into the dark part of a box, in educating it to turn back as soon as it reached the edge of the darkened region, without waiting for the shock, and Turner (729) obtained similar results. When an instinct is thus completely suppressed by “punishment, the conscious accompaniment of this modification in behavior is probably simply a change in the af- \fective tone of the situation. Instead of being pleasant, ‘it becomes unpleasant. In a human being, memory ideas | might accompany the process: a human pike, for instance, | might at the sight of a minnow recall clearly the bump on \ the nose and his consequent humiliation. But we can fexplain the pike’s behavior just as well if, in accordance with Lloyd Morgan’s canon, we assume merely that the sight of a minnow has become unpleasant to him: he has ost his taste for minnows. Punishment has been the means in many cases of training animals to manifest their ability to discriminate beeen stimuli. The desired end is of course to attach the negative reaction to those features in which the “‘wrong”’ timulus differs from the ‘‘right”’ stimulus. For instance, an animal is being taught to choose a light rather than a dark passage, the two openings being side by side: when he enters a dark passage he gets an electric shock. It will be natural for him at first to attach the withdrawing reaction consequent on the electric shock to the sight of the whole ‘apparatus. Whether he will shrink back from it or rush indiscriminately into either of the passages depends on the relative prepotency of his impulse to enter the passages and his impulse to withdraw from injury: in either

case he makes no discrimination. The discrimination occurs when the withdrawing reaction attaches itself to the feature which distinguishes the dark passage from the rest of the apparatus, namely, its darkness. It is probable that in many cases the animal does not deliberately compare the light with the dark passage, but merely learns to | distinguish the passage to be avoided from the rest of the | situation at large. We should expect this to be the case ‘ where punishment is the only method of training used: | the case would not be one of “white preferred to black,” / but of ‘‘anything rather than black.”

The strength of the punishment applied is of course an influential factor in the learning. Obviously it depends not merely on the strength of the punishing stimulus, but | on the sensitiveness of the punished animal. Yerkes was the first experimenter to employ the electric shock as a means of training animals. He used it on the frog (805), which he was trying to educate to make a turning to the left rather than to the right: the frog showed a discouraging tendency to sit motionless for long periods of time, and so Yerkes placed electric wires on the floor, to induce by a mild shock greater activity. In his work on the dancing mouse (820), he substituted the giving of electric punishments in the case of wrong choices, for the older method of rewarding an animal’s right choices, and one of the advantages claimed for this method was that it seems to allow an exact measurement of the strength of the stimulus, whereas a reward, such as food, varies in strength with the animal’s physiological condition. But the effect of an electric shock } too varies with the temporary physiological state of the animal, and with its general individual sensibility. Yerkes (827) carried out some interesting experiments on the relation of the strength of the punishment to the difficulty

of the discrimination required of the animal. A superficial consideration of the situation might assume that if one wants to teach an animal a difficult discrimination, such as that between two slightly different shades of gray, one ought to supply a stronger punishment stimulus than would be necessary to teach it an easy discrimination, such as that between black and white. The results with the : dancing mouse showed on the contrary that weaker pun- | ishments were more effective in the learning of hard dis- | criminations ; stronger punishments in that of easy dis- \criminations. The same rule was found by Dodson to hold for cats (186 a); the hardest discriminations were acquired by kittens in 82.5 trials with a moderate stimulus, but 107.5 trials were required on the average with a strong stimulus. Indications of a similar relation were found by Cole (136) in the learning of chicks.

The negative reaction is not the only one which may show sufficient prepotency to cause the dropping off of other responses. The feeding reaction, or any other innate response, may serve: thus reward as well as punishment is a method of training. The taming of an animal by kind treatment illustrates both the simple dropping off of useless movements, the getting used to a situation, and the substitution of movements more valuable to the animal; the tamed creature on the one hand learns to rest quietly in the presence of its tamer, instead of displaying alarm, and on the other hand to come for food or follow for companionship. A very pretty illustration of the overcoming of an innate response to light by the response to feeding was obtained by Wodsedalek (795) on immature mayflies. These insects have an innate tendency to avoid light and to remain under stones in the water. By regularly feeding them on the upper surface of a stone the

experimenter was able wholly to overcome this reaction, especially with one gifted individual. After two months of training, ‘‘all that was necessary to bring the specimen up when it had disappeared from sight was to slightly jar the dish or the table on which the dish was located, and the insect would quickly come up to the upper side of the rock and make for its feeding place.”,, Here again, the conscious aspect of the learning is probably a reversal of| the emotional tone of the situation: originally unpleasant, it has become pleasant. Where the method of reward is’ used to train animals in discriminating stimuli, the influence of the reward is combined with that of the tendency to drop off useless movements. Cole’s raccoons learned not only to climb up when the food signal was given, but to stay down when the no-food signal appeared. The rabbits studied by Miss Abbott and the writer (756) were taught to push at a door carrying a piece of red paper, and to refrain from pushing at a door carrying gray paper. The original stimulus for the pushing was the odor of food which was in the compartments behind both doors. The “gray” door was always bolted on the inside, so that pushing against it was in vain; the “red” door opened freely so that the rabbits could get at the food. The actual} securing of the food acted, along with the smell of it, to suppress all useless hesitations on the part of the animals and to make them more inclined to push the doors at once; the gray stimulus acquired a tendency to lose its motor effect because the movements to which it gave rise were useless.

So-called ‘‘puzzle-box” experiments also depend for their training effect upon the combined tendencies to the survival, through their prepotency, of movements resulting in the satisfaction of an instinct, and to the dropping ( of of useless movements. The method has been tried with birds, rats, squirrels, cats, dogs, raccoons, porcupines, and monkeys. Thorndike, its originator, made some experiments of this type on chicks confined in pens from which they could be released by pecking at a string or some such object (704). Porter tested English sparrows with boxes containing food, which could be en-

Fic. 12. — Puzzle box used in Porter’s work on birds; AB, one method of attaching string to latch; C, a second method. In the first, the loop at B had to be pulled; in the second, the string had to be pushed in. tered by pulling a string fastened to a latch, or by pushing the string into the wire netting with which one side of the box was covered (Fig. 19 The sparrows learned very quickly; one of them by the tenth test had left out all unnecessary movements (610). In later experiments a cowbird and a pigeon also learned to open a similar box. Before beginning the test the birds were accustomed to being fed in the box with the door open. Their first success in opening the door lay in accidentally clawing or pecking at the proper point, and in later trials the action