282 ABRIDGMENT OF LEARNING PROCESS again by the introduction of another animal. Chickens hasten more rapidly through the maze if they can hear their fellows pecking on the outside, etc.
Putting the animal through the act. — A reference to the literature will show that positive results are- being accumu- lated by the use of this method. The theoretical envisage- ment of the problem involved in the interpretation of these results is difficult. All we can do, apparently, is to admit the general fact and to wait for more extended w^ork to bring analysis. If putting the most motile organ (paw, beak, etc.) of an animal slowly through the various parts of an act brought a hastening of the act as a whole, inter- pretation would at least be theoretically possible. But with the exception of Cole, in his work upon the raccoon, all have failed to get positive results from this method. On the other hand, it is possible to get both the rat and the raccoon to climb to the top of a box and drop through a trap to the floor of the box (which is quite foreign to their instinctive organization) after they have been passively dropped through the trap on various occasions. It would seem that in this method we have an additional (i.e., to the perseverance method) way of establishing a simple co- ordination in an animal. Yet we are not willing to admi^ this until there has been made a very thorough analysis of the factors which enter into it (p. 293).
Presentation of experimental results: introduction. — From these preliminary statements concerning the general problem of higher forms of learning it will be seen that it is impossible to present the results which have been obtained in any thoroughly systematic and unitary way. The technique of making the tests has been as varied as the presuppositions held by the different investigators. Ac- cordingly the results will be presented in some detail.
A. Primates. — Thorndike concludes in general that learning in the monkey (Cehus) cannot be hastened by the methods under consideration. He found that opening the door of a given box for the monkeys had no effect upon the rapidity with which they learned to open it. Thorndike 's usual method was to ' ' attract the attention ' ' of the monkey EXPERIMENTAL RESULTS 283 or to wait until the animal was looking at tlie experimenter, and then to open the box slowly, emphasizing the move- ments. The monkey would then be given a trial without tuition; or the experimenter would repeat his act several times before allowing the animal to try the box. All ex- periments of this type failed to yield positive results. They profited just as little by the tuition of other trained ani- mals. One animal was allowed to open the box by the perseverance method. This animal was then used as the imitatee. The imitatee worked with the imitators near at hand. Negative results were again obtained. Monkeys fail to learn, according to the same writer, by being put through the movements. Thorndike made several such tests upon the monkeys. One animal was carried to the top of a table and fed there, and then tested to see if it would go to the table of its own accord. This was repeated ten times, and then the animal was released. In no case did this monkey or any other give an imitative response. Watson's experiments were made upon monkeys (M. rhesus) which were quite tame. Two of the monkeys were quite friendly with each other. J was an adult monkey and B a young one. In a short time an attachment sprang up between these monkeys which persisted for several years.
B was wild and restless and became excited when J was away from him. J on his part would run to B on certain cries and put his fore-paws around B's neck. B would nestle up to J and clasp him around the chest. If B did not become placid under this treatment, J would rock him from side to side and give out a little companionable chat- ter. Under these conditions, it is quite evident that J's reactions influenced B 's greatly. When J went to one part of the cage B followed. If, while sitting on the shelf, a pan of water or bowl of milk was placed upon the ground, B would not come down to drink if J did not precede him. J formed the habit of jumping on to the experimenter's shoulder when he entered the cage and called. B formed the same habit, but if J for any reason refused to come, B refused also. If B were loose in the room when the experi- inenter passed with J on his shoulder, B would run to the 284 ABRIDGMENT OF LEARNING PROCESS experimenter immediately. But, on the other hand, if J was not on his shoulder, B paid not the slightest attention to the experimenter's commands. At times J and B were left loose in the animal room. Sometimes J was refractory and did not care to go into the cage. B would never go in unless J did. It was necessary to threaten him at times with a long stick. J would walk around the cage two or three times before entering it, but always just out of reach. B trailed J and when the latter finally darted through the door of the cage B darted in also. It was almost impossible to force B in unless J had preceded him. Several types of problems were given the animals in the work upon imita- tion. (1) Drawing in food with a rake, animal to imitate the experimenter; (2) drawing in food with a cloth, animal to imitate the experimenter; (3) obtaining food from bottom of bottle by the use of a fork, animal to imitate the experimenter; (4) pushing out food from middle of long glass cylinder by means of stick, animal to imitate experi- menter; (5) manipulation of old-fashioned latch, animal to imitate experimenter; (6) box with a door in top not held in place by any fastening, animal to pull open door by means of a handle, B to imitate J; (7) box with door in top held in place by push button, B to imitate J. The monkeys very quickly form by their own unaided efforts such habits of manipulation, but so far as our observations went they were entirely uninfluenced by tuition. In the simple prob- lems, such as pushing out food from a glass cylinder, pull- ing in food with a rake or cloth, etc., we failed wholly in obtaining positive results. When the animal is placed in such situations it at once begins to strain at its tether and to reach out after food with its paw. It would often pick up the stick or cloth, bite it, and then discard it, but it would never attempt to go through with the act of push- ing out the cloth or stick, throwing it over the object, then pulling it in. In other words, our conclusions exactly har- monize with those of Thorndike and are entirely against Hobhouse's type of results. We did find evidence for certain types of reaction which are generally included under the head of instinctive imitation. With Kinnaman we EXPERIMENTAL RESULTS 285 found that if one monkey discovered a hole and peeked into it, another would generally push him aside and peek into it in his turn. This was observed several times. More re- cently we have confirmed these observations many times in the case of a mother and baby. One day the baby left the mother 's arms and peered under a sill at a brass lock which could be seen from his viewpoint but not from hers. He pushed his arm through and attempted to pull the lock forward. The mother left off eating, came and adopted the same position and reached her arm through and attempted to grasp the lock. The two acts were identical in techniciue. We found also that certain acts which resulted in a sound would be rhythmically repeated. In one case one of the monkeys (Cehus) filched a large tablespoon. He was found standing the spoon on one end and then immediately releasing it. Dropping the spoon seemed not to be acci- dental but actually a part of the act. He repeated this act 15 times in unvarying order and, so far as could be judged, at definite intervals of time. This corresponds, in our opinion, very closely to the child's act of repeatedly ham- mering its spoon against the table. Hammering with a nut or any small, hard, preferably round, object was another favorite trick of the Cehus monkey. In one case the animal tapped with a small hickory nut 150 times in one half hour. In the forest this is possibly connected with some kind of food-getting process, although we have no assurance of this. Rhesus J also gave evidence of a similar type of action. Occasionally he would sit and play with a small hard object, let it slip through his fingers to the board on which he sat, pick it up again, and drop it. One instinctive act could at times be called out in J, that of flea catching. As is well known, this is probably the most frequent act indulged in during their quieter moments. The smacking sound used by the monkeys in this act is easily imitated. On one occasion J allowed the experi- menter to hunt fleas over him. After performing this act J perched upon his shoulder but made no attempt to pick him. The experimenter then held up the hairy portion of his arm to J but he made no effort to hunt for fleas! The 286 ABRIDGMENT OF LEARNING PROCESS smacking sound was then made with the lips. J immedi- ately made the sound in turn and began searching the arm, and then proceeded to the experimenter's neck (the experi- menter was wearing a rubber cap which extended down over his ears). Haggerty worked for several months on the rhesus and Cehus monkeys. The animals were placed in the presence of simple mechanical devices the manipula- tion of which opened doors, disclosed openings, or dropped food into the cage. Each animal was given a fair oppor- tunity to imitate in a series of preliminary trials, — usually 15 in number, each 15 minutes in length and given on suc- cessive days. In almost every case the animal had either solved the problem by the end of that time or had lost in- terest in it. Imitation experiments were then begun upon the animals which had failed. The trained animal was allowed to perform in the presence of the imitator. After this the imitator was given an opportunity to try the mechanism. He was allowed to work 10 minutes, and longer if he seemed on the point of making the successful move- ments. If he failed the test was repeated. He was not considered as having failed to imitate until 100 opportuni- ties for profiting by tuition had been given. Under these conditions Haggerty obtained 16 cases of imitation, 3 of which were immediate, and 5 cases of partiallj^ success- ful imitation. Of the 11 animals used hy him but 2 failed to exhibit imitative behavior in one test or another. Hag- gerty mentions the fact that animals imitate best when not too thoroughly accustomed to one another. He states that these cases are cases of imitation of the inferential type. We are not able to agree with him. In the first place, his method of allowing the animals to have 100 chances for imitation before considering tuition without effect must alwaj^s give ambiguous results. "With an animal as agile and as varied in his interests as the monkey, it is never possible to tell when a burst of activity will lead to the solution of the problem by the perseverance method. On the other hand, it is possible to assume that the only effect of the act of the imitatee was to set off two congenital forms of response on the part of the imitator. The one was the EXPERIMENTAL RESULTS 287 following instinct, — i.e., to go to the place where a group of monkeys had just been and to peer into the same holes that other monkeys had just peered into, etc.; and the other the tendency to attack moving objects first. The general perseverance method of the animal took care of the remain- ing factors. Kinnaman reports instances where he observed that the acts of the male monkey {M. rhesus) influenced the responses of the female. In one case where food was to be had by pulling out a plug, and in another where food could be obtained by bearing down upon a lever, the female had had opportunity to learn without tuition but had failed. The male had learned by his own efforts. He was allowed to perform in front of the female. She went at once in each case, after seeing the male get the food, and operated the mechanisms, and then repeated the acts later several times. Witmer was not very successful in getting Peter, the chimpanzee, to imitate the trainer in fitting the form blocks into their appropriate holes. He states that better success was obtained with him in the imitation of the letter W, which was placed upon a blackboard. Witmer also cites the case of a monkey {M. cynomolgus) in opening a door. He maintains that this door was opened for the first time by the monkey. There is nothing in his paper that would show that he knew anything about the previous history of the monkey. Opening a door is one of the easiest things that a monkey does. Time and time again we have had them unhook doors, twist off wires which were wrapped around doors, break wires, take a round knob and turn it with the hind-feet while holding on to a nail in the wall with the fore-feet. All of these acts were readily learned without tuition and in the absence of the experimenter. It is very easy to be deceived about the accomplishments of an animal not.under daily observation. Shepherd repeated several of the tests made by Watson and by Hobhouse. In one of his imitation tests he used a glass tube 15 inches in length and | of an inch in diameter fitted with a plunger which could be easily taken out and inserted. All failed to imitate. The animals would take up the plunger, throw it down, pick it up, insert it in the mouth, bite it, and then 288 ABRIDGMENT OF LEARNING PROCESS again discard it. They were also tried with a T-rake. Nearly all failed. None pushed out the rake and hooked the prong over the food. Monkeys 3 and 7 learned to push out the rake and with much slashing about of the T end to succeed in drawing in the food. The act was very awkward for them. One of the monkeys would many times knock the bit of food away. The best that can be said is that the animal learned to imitate the experimenter in pushing the rake out, but only awkwardly in hooking the rake around the food and pulling it in. These coor- dinations were never smoothly executed by the two monkeys. There is some slight evidence that one of them learned something from the tuition of the other. In another type of test positive results of a fairly satisfactory kind were obtained. A banana was suspended by a string from a horizontal rod attached to the wall of the room. The banana was entirely out of reach of the monkey unless he would push a horizontally placed pole (pivoted at one end and supported by a railing at the other) under the banana. Having done this he was able to mount the pole from the ground and to reach the banana from the pole. While in general the monkeys learned this problem by the perseverance method, there were certain indications that imitation was involved. In the case of monkeys Hobhouse cited the behavior of the chimpanzee P which had learned previous to the tests under consideration to throw his rake out and over the food that was out of his reach and then clumsily to roll it in. Hobhouse taught P to substitute a stick with a crook at one end for the rake. Food was placed in a light box. The chimpanzee learned in time to hook the end of the stick into the box and pull it in. The animal learned rapidly to use a short stick to bring in a longer one — the latter enabling him to. obtain food which was beyond the reach of the short rod. Hobhouse tried to teach the chimpanzee to throw a rope with a noose in it over the box and bring in the food in this way, but even when he got the noose around the box he did not complete the response by pulling it in. The M. rhesus mon- keys were much slower in learning to use the stick. They EXPERIMENTAL RESULTS 289 did not succeed apparently any better than those used in the author's experiments. Hobhouse's M. rhesus did not at first learn to use the short stick to pull in the longer one. He later learned to use a stick quite frequently and when it was out of reach even to use substitutes for it, such as a child's skipping rope, wire, cord, etc. Hobhouse reports an interesting act in the M. rhesus: food was placed on a table just out of reach of the monkey. However, the tether was long enough for him to reach the food provided a chair or box was pushed between him and the table. He rapidly learned to push the chair into position and to mount it.
B. Cats. — Berry states that the Manx cat is very much influenced by the behavior of its associates. In learning a problem box a trained animal was used as the imitatee. In one case it was shown that cat M refused to turn a but- ton which would have released her until cat X had turned it several times in front of her. In rolling a ball into a hole cat Y refused to do it until the trained animal had performed in front of him. The presence of imitation was found in such acts as pulling a knotted string which would release a door. Instinctive imitations such as going to the same place and looking down the same hole, scratching at the same place, and so on, are also reported by Berry. Thorndike failed to get cats to imitate. Trained animals were allowed to perform in front of the untrained animals. Several simple experiments like the following failed to yield positive results: The cats were placed in a large box with a door in the top; the experimenter went to the box and opened the door to get the animals out. The first cat up was always taken out first and fed before the others. Usually they climbed up to the door on the ap- proach of the experimenter. Of 3 recently purchased ani- mals 1 had not acquired the habit by the end of the tenth day. Even though a piece of fish were held up to him he would not climb up. It is reasonable to suppose that imitation, if at all prominent in their lives, would have overcome this sluggishness. Thorndike then put 2 trained cats in the cage with this animal and had them climb up 290 ABEIDGMENT OF LEARNING PROCESS 80 times before his eyes and get the food. No positive re- sults were obtained. In one other interesting test they again failed. All animals of a certain group were taught to climb to the top of a large cage on the approach of the experimenter with food. Two animals, then, were trained to go up on a signal. One of these trained animals was put into the cage with the untrained animals. On the signal the trained cat would go up, but the untrained animals would not go up* until the approach of the experimenter. This association was not very quickly formed, although it was learned in time by the perseverance method. Warren re- ports a case of delayed imitation in the cat. One of the two pets formed, through efforts of its own, the habit of climbing into the experimenter 's lap, then on to his shoulder and out over his outstretched arm to a piece of meat held on a fork. The companion of this cat made no attempt to perform the trick, although watching the trained animal feed in this way daily. This second cat had often been coaxed to make the effort. One day, however, (after the owner had been absent for some time) after watching the trained animal feed, it suddenly sprang into the experimenter's lap, walked out over his arm, and seized the meat. After this it became the sturdy rival of the cat which had first learned the trick.
C. Dogs. — Thorndike has made similar tests upon the dog. After the dogs had failed to open a given box, the experimenter went through the act. The dogs failed in all cases to profit by such tuition. One dog had been taught to jump up on a box and " beg " when a piece of meat was held above his head. He did this 110 times in the presence of an untrained animal. The untrained ani- mal never learned to imitate. Putting the animal through the act proved just as unavailing as in the case of monkeys. In such experiments the animal was left at its task for a few minutes, and if it failed to work the mechanism the experimenter would place his hand in the cage, take the dog's paw and force it through the act of pulling the string or working the lever, as the case might be. This he would repeat again and again. At the end of 10 to 15 EXPERIMENTAL RESULTS 291 minutes he would test the animal alone. None of the ani- mals failing to learn by its own unaided efforts ever learned by being " put through."
D. Raccoons. — On the question as to whether raccoons imitate there is much difference of opinion. Cole says that they imitated his movements in pressing down the lever of a card displayer (which finally resulted in their obtain- ing food). Sometimes they apparently worked the lever as the experimenter had done and at others at least at- tempted to work it in that way. After a few days they began to watch the experimenter's hand very closely. He admits that they do not imitate their mates. He finds further that the raccoons learn by being '^ put through." He tried Thorndike's experiment of dropping the animals into the experiment box. The animal was picked up by the back of its neck and dropped into the box. It had to open the door to get out and obtain a bit of food. At the end of 33 to 57 trials the animals learned to turn and go back into the box if the door was left open. If they found the door closed the raccoons would climb up the sides of the box and drop in through the hole at the top. In the case of these animals there is some evidence that they learn more quickly if passively ' ' put through ' ' the movements of working a simple mechanism. The main effect appears im- mediately after they are ' ' put through ' ' the act. The effect is very transient. An animal never failed to work the mechanism if given a " put through " trial at first. He otherwise often failed in the early stages. Cole states that they can be made to learn an act by tuition after having failed to learn it without tuition. The animals did not al- ways duplicate the act exactly, although they could be taught to do even this. Shepherd failed to get such satis- factory evidence of imitation in the raccoon. His animals had been trained to solve problems involving the manipula- tion of levers, etc., for about 3 months. His imitation apparatus consisted of an inclined plane of poultry netting 25 cm. wide supported at one end by a box. From the box the plane extended diagonally across the room to a platform. The raccoons could mount the box from the floor of the 292 ABKIDGMENT OF LEARNING PROCESS room. At a given signal the trained raccoon would mount the box and walk up the plane to the platform where he was fed. This animal was the imitatee. Two others were used as imitators. Several experiments failed to show a more rapid learning on the part of the imitators than would be called for by the assumption that their learning was of the perseverance type. Davis reports one case of doubtful imitation in the raccoon. For the most part he failed to find any trace of it. Two experiments are cited: A blind coon was taught to climb to a platform in the middle of the room. Upon this he would stand and beg for food. Another raccoon was taught to climb up' with the blind animal, but although this animal had seen the blind coon stand up and beg for food many times it w^ould never, on its own account, imitate the act. One animal was taught to jump for food up to 3 feet in height. He was always in the midst of the pack when he performed this act, yet none of the other animals ever attempted to imitate it.