E. Rodents. — Berry states that the white rat will imitate simple actions. If one rat begins digging another one will come up and begin to dig also. In other experi- ments less instinctive acts were said to have been imitated. It is extremely questionable whether the rat has sufficiently developed visual receptors for such a complex stimulus as the performance of another animal to lead to such com- plicated motor results. Such results would be very dif- ficult to account for when we consider that it has been shown that the rat learns with great difficulty to respond to even large differences in size, pattern, and intensity of visual stimuli. Dr. Florence Richardson in her work at the University of Chicago was not able to obtain positive results on rats when they were tested in an apparatus as nearly like Berry's as could be made from his description. Imitation of the following type has been admitted by every one since the work of Small appeared. Hunter finds that the white rat can learn by being '' put through " an act. As a part of the daily routine of an experiment upon a group of rats it was found necessary to deposit each ani- mal, after completing his quota of work, in a small box.
EXPEEIMENTAL RESULTS 293 This was accomplished by lifting the animal up and drop- ping it through a hole in the top of a box resting on the table which held the living cage. The animal remained in the box until all the members of the group had been worked with, after which the group as a whole was removed to the living cage and there fed. On the two hundred and fourth day after experiments had begun the door of the living cage was left open by accident. Two of the 5 rats climbed to the top of the small box and dropped to the floor of their own accord and remained there, as was their custom during the regular routine of experimentation. A regular set of tests was then initiated. Several results of this character were obtained." Yerkes failed to get imi- tation in the dancing mouse. He allowed a trained animal to escape from a box by climbing up a ladder. This trained animal performed before untrained animals many times. The untrained animals failed to profit by the copy. A similar condition of affairs resulted when discrimination tests were tried. The animals responded individually. Seldom would one even follow the other. Again, some of his animals were taught to pull and others to push in a door of an experiment box. Those which had learned to push out doors did not learn to imitate those which had learned to pull in doors in situations w^hich demanded the latter type of response. Although abundant opportunity was given in a variety of acts, no imitation was found. This has been confirmed by other observers on different varieties of mice. Allen states that the guinea pig can be ^' tolled " through certain movements (thus supplying ^ It is possible, but probably unwise, to interpret this act in a rather simple way. After the animals are dropped into the box a habit of remaining there is formed. The floor of the interior of the box often reacted to may, whenever sensed, lead to the old response of jumping down upon it. The rats, when let out inadvertently, do as all rats do, run over every available inch of surface. On arriving at the top of the box the stimulus of the floors and sides leads to the old response. Rats often climb over the interior of a box and over its cover, jumping at times to the floor. There are many obvious ways of controlling this response, by the use of blind and anosmie animals, by putting the animal through a door in the side, etc. Until some one does this, we shall not know what factors are involved.
294 ABRIDGMENT OF LEARNING PROCESS its own motor impulse) which has the effect of hastening the formation of the habit.
F. Birds. — Thorndike gave his chicks the chance of imitating the act of pecking at a door, jumping up and with the neck pulling down a string, stepping on a plat- form, pecking at a tack, etc. In no case was imitation found. Porter offers a new criterion of imitation, as fol- lows: Bird No. 1 is allowed to open a problem box by the perseverance method. Bird No. 2 which has not learned to open it, or else has a different and possibly a poorer method of opening it, is allowed to work upon the problem with bird No. 1.^ If either bird changes its method, dis- carding a poor one for a better one, or if the untrained animal should suddenly pick up the method, imitation is present. In all cases, except where the tests were made upon blue jays, Porter found positive evidence of imita- tion. The birds giving positive results were English spar- rows, cow birds, juncoes, white-throated and field sparrows, song sparrows, Baltimore orioles and crows. Birds of the same species imitate one another more readily than they do birds of a different species. Baldwin in personal con- versation mentioned a case of delayed imitation in the parrot. He tried to teach this parrot to sing '' Over the Fence Is Out." The bird would never attempt to repeat the words while the experimenter was present. One morn- ing, though, as he came down the steps, the bird started in with the melody.^ The observations of Scott and of Con- radi, showing that the type of song which develops in a given bird is largely dependent upon the nature of the songs of the other birds confined in its neighborhood, have already been mentioned in another connection (p. 142). Lashley has recently confirmed this work and ex- ♦ The two birds are likely to fight. Porter supposes that by this process he makes the stimulus to get at the food or to work at the problem far greater than by the ordinary method; e.g., he states that " rivalry " is a strong motive.
^ He has since published a similar observation: Baldwin, J. Mark. "Deferred Imitation in West-African Gray Parrots." Extrait du IX® Congres international de Zoologie, tenu a Monaco du 25 au 30 SUMMARY 295 tended it. He finds that when an Amazon parrot (Chrysotis)f which does not talk, is confined with a bird which does, the untrained animal begins to repeat at first very indistinctly the words of the trained parrot. These gradually become more definite. The untrained animal in this case profited much more rapidly from the tuition of its fellow than from that of the experimenter. This same investigator Ifas tested to what extent a well-trained parrot can reproduce articulate and inarticulate sounds. This bird possesses 50 to 100 distinctly articulated words which can be called out by appropriate stimuli. In addition, he sings, whistles, barks, mews, cackles, and laughs. When tested on ability to respond in kind to such sounds, he gave 30 whistling tones to 30 whistling stimuli produced by the experimenter's lips and by blowing a small metal whistle. He responded to 31 singing tones from violin, cello, piano, and voice 29 times by singing tones and 2 times by whis- tling tones. The bird was able to reproduce varied types of sounds as is shown in the following table: Stimulus Times given Response Times given Guttural speech 10 Guttural muttering.. 10 Whistling 10 Whistling 10 Whispering 10 Whispering 10 Cough 10 Cough 10 Whispering If Further tests tend to show conclusively that there is actual reproduction of vibration frequency. The bird was not able to reproduce a given succession of notes (not in his repertoire) given by the experimenter.
Summary. — It will be seen from the above summary of the work on imitation that the results are not harmoni- ous. Indeed, in some cases two authors working upon the same species have arrived at exactly opposite conclusions. On the whole, however, enough positive results have been obtained to show that in some way the process of learning is modified when certain forms of tuition are used. If further work were conducted along lines which are not an- thropomorphic— if the experimenters would confine their 296 ABRIDGMENT OF LEARNING PROCESS work to determining just what the animals do without being concerned about the fact that they may be acting as human beings would under similar circumstances — the control of behavior by such methods in both human and animal work would more speedily be realized.
BIBLIOGRAPHY Allen, Jessie, " The Associative Processes of the Guinea Pig," Jour. Comp. Neu. and Psych., 1904, XIV, 293. Berry, C. S., " An Experimental Study of Imitation in Cats," Jour.
Comp. Neu. and Psych., 1908, XVIII, 1.
"Imitative Tendency of White Rats," ibid., 1906, Cole, L. W., " Concerning the Intelligence of Raccoons," Jour. Comp.
Davis, H. B., " The Raccoon: A Study in Animal Intelligence," Am.
Jour. Psych., 1907, XVIII, 447. Haggerty, M. E., " Imitation in Monkeys," Jour. Comp. Neu. and HoBHOUSE, L. T., Mind in Evolution. New York, Macmillan, 1901. Hunter, W. S., "A Note on the Behavior of the White Rat," Jour.
KiNNAMAN, A. J., " Mental Life of Two Macacus rhesus Monkeys in Captivity," Am. Jour. Psych., 1902, XIII, 98, 173. Lashley, K. S., " Reproduction of Inarticulate Sounds in the Par- rot," Jour. Animal Beh., 1913, III, 361. Morgan, C. L., Animal Behavior. London, E. Arnold, 1900. Porter. J. P., "Intelligence and Imitation in Birds; a Criterion of Imitation," Am. Jour. Psych., 1910, XXI, 1. Richardson, Florence, " A Study of Sensory Control in the White Rat." Psychological Monograph, Ser. No. 48. Shepherd. W. T., " Imitation in Raccoons," Am. Jour. Psych., " Some Mental Processes of the Rhesus Monkey," Psychological Monograph, Ser. No. 52.
Thorndike. E. L., Animal Intelligence. New York, Macmillan, 1911. Warren, E. R., "Delayed Imitation in a Cat," Jour. Animal Beh., Watson, J. B., " Imitation in Monkeys," Psych. Bull., 1908, V, 165. WiTMER, Lightner, "A Moukcy with a Mind," Psych. Clinic, 1909, Yerkes, R. M., The Dancing Mouse. New York, Macmillan, 1907.
CHAPTER IX THE LIMITS OF TRAINING IN ANIMALS Introduction. — Some gifted animals. — (1) Clever Hans. — (2) The horses of Elberfeld. — Claparede's examination; results of the first set of observations. — Later tests by Claparede. — Tests on morning of the 26th of March. — ^Afternoon of March 26th. — March 28th, morning and afternoon. — Tables of results. — (3) Peter. — (4) Don. — (5) Jasper. — Summary, Introduction. — From time to time there loom above the level of the behaviorist's horizon reports of individual animals or groups of animals that affirm that at last a prodigy has been found which possesses something special in the way of a behavior equipment. Such animals have never been discovered in the laboratories devoted to the study of animal behavior. That the highly gifted animal should not thus appear in the laboratories is understand- able when one considers' that investigations have heretofore been rather narrowly concerned with the instincts and the sensory and motor habits which appear in a laboratory environment. The gifted animal has usually been de- veloped by the amateur. Through the animal's contact with its owner and with other animals there arise highly complex modes of response. The training methods are not controlled and no scientific attempts are made to analyze the exact nature of the stimulus to which the animal re- sponds. From an anthropomorphic standpoint the animal apparently is reacting as a human being would act under the same circumstances. In a short time the doings of the animal get noised abroad and it becomes necessary for some trained investigator to step in and reduce the chaos to some semblance of order (interpretation of acts where there is ignorance of training methods is not an easy task). "Within recent years several animals have appeared which have caused a certain amount of consternation among 298 THE LIMITS OF TRAINING IN ANIMALS investigators who have gone to examine into the phenomena. The situation is exceeding^ like that which appears in the investigation of so-called occult phenomena.^ Some new medium arises. Some prominent man visits the medium and becomes mystified. A scientific man, usually a physicist, is selected to investigate her. The physicist re- ports an elaborate series of tests which shows that the medium is not using concealed wires, magnets, mirrors, or other physical equipment. The physicist himself may be- come " convinced." The mystery grows. Finally the psychologist makes the test and finds some simple trick which will account for the phenomenon. The number of such alleged occult phenomena from telepathy to spiritual- istic converse, which have been investigated and found want- ing in scientific or philosophical interest, has been so nu- merous that now it is very difficult to get a psychologist who values his reputation to undertake such an investigation. In the same way, when a wonderful animal appears, zoologists, botanists, and physiologists are hastened to the scene. A commission is appointed and the mystery deep- ens.. Usually when some man who is familiar with the methods of training animals and ' with animals ' methods of responding is found, the explanation, while not necessarily simple, smacks not at all of the mysterious. This situation in the past, with regard to the animal world, has not been wholly without beneficial effect. In the first place, it has brought the behaviorist face to face with the fact that there may be depths in the animal to which he has not descended and cannot descend except by adopting a part, at least, of the technique of the amateur, viz., that of living a large part of his time with the animal and com- plicating the methods of training (as is done for the child). That this meritorious effect has been produced is shown by the fact that there are investigators in behavior who are willing to devote a large number of years to the study of a single animal or at most a small group of such ani- mals. The German station on the Canary Isles is a case ^ Attention is called to President Sanford's " Discussion of Animal Prodigies," Amer. Jour. Psych., 1914, p. 1.
GIFTED ANIMALS 299 in point. Along with certain biological investigations will go detailed studies on the behavior of the anthropoid apes. It is certain that this tendency to specialize on one or at most a few animals will become more common. Since the higher anthropoids are nearest to man in their equipment and since their tenure of life is long under natural condi- tions it is to be expected that students will undertake work upon them with greater willingness than upon lower forms. To do such work effectively there is great need of an anthropoid ape station open to American students. The second generally good effect the study of the gifted ani- mals has had upon the behavior work comes from the fact that in making an analysis it is often necessary to under- take a study of the sensitivity of their receptors, field of vision, limen of sensitivity for moving objects, etc.
Some gifted animals. — On account of its bearing upon the limits to which training in animals may be carried, we wish to examine into the behavior of some of the better known gifted animals. We shall discuss in order (1) the horse, Clever Hans, (2) the horses of Elberfeld, (3) the chimpanzee Peter, (4) the talking dog Don, and (5) the dog Jasper. In considering their behavior it seems well to bear in mind that all behavior, human and animal, is analyzahle in terms of stimulus and response, and that the only difference between man and animal upon this assump- tion would be in the complexity of behavior. The funda- mental difference between man and animal from our point of view lies in the fact that the human being can form habits in the throat (and other bodily language habits), — neglecting his finer sensory-motor equipment. All other habits pale into insignificance when contrasted with these. "While this at present is affirmed on the basis purely of theory, we feel that the assumption has a good many points in its favor (p. 322).
Clever Hans. — In 1890 Mr. von Osten, a gentleman of Berlin with some mathematical training, noticed what he considered remarkable behavior in a horse which he then possessed. "When that horse died he purchased another (1900) and began training him about a year later. This 300 THE LIMITS OF TRAINING IN ANIMALS horse, Hans II, was destined to arouse an enormous amount of interest both in popular and in scientific circles.
Mr. von Osten's method of instructing the animal was very- simple. He began by teaching the horse to respond properly to the different words for spatial localization, such as left, right, above, below, etc. Simple arithmetical problems were next attempted. He would place one, two, three, etc., objects upon a table, kneel down and take the fore-foot of the animal ana make him tap once for each object. In a remarkably short time Hans learned to "count" and to perform simple additions and multiplications. But these attainments did not satisfy von Osten. He desired to give Hans a wider medium for the expression of his talents. Accordingly von Osten arranged for him a conventional alphabet. Having mastered this the horse was able to read and to answer simple questions. At the height of his career Hans showed the following scientific accomplishments (summarized from Pfungst): he had mastered the cardinal numbers from 1 to 100 and the ordinals to 10. Upon request he would count objects of all sorts, the persons present, even to distinctions of sex; then hats, umbrellas, and eyeglasses. All forms of simple arithmetical problems involving addition, subtraction, mul- tiplication, and division. He could do fractions, changing them first into decimals. E.g., "How much is f -[- Vo? " Answer, y^^ (tap- ping first the numerator and then the denominator). Or again: "I have a number in mind. I subtract 9 and have 3 as a remainder. What is the number?" (12). "What are the factors of 28?" Whereupon Hans taps consecutively 2, 4, 7, 14, 28. He was able to read German readily, whether written or printed. If a series of placards Avith written words was placed before the horse he would step up and point with his nose to any of the words required of him. With his alphabet he would answer simple questions, e.g.: "What is the woman holding in her hand?" Hans spelled without hesitation " Schirm " (parasol). Furthermore, his memory was ex- cellent. He carried the entire yearly calendar in his head. He could give not only the date of each day without having been previously taught, but also the date of any day one might mention. He could tell the time to the minute. His musical ability was especially well developed. He had absolute pitch memory and a feeling for intervals. He could analyze compound clangs without difficulty, and could inform his experimenters as to what changes should be made in a discordant clang in order to produce consonance.
Naturally Hans' fame spread and different people visited him. These at first were connoisseurs of horses, officers, who — astonished — told about Hans' accomplishments. Soon the public at large became interested and von Osten's court-yard became too small to accommodate the curious. The newspapers discussed it as the " Case of Hans." Vio- lent polemics followed between those who believed in the CLEVER HANS 301 animars intelligence and those who suspected some trick. These latter did not agree among themselves, — some talked of optical signals, others of acoustical signals, which von Osten gave his horse. Others supposed that electrical wires placed under the pavement of the court allowed the master to communicate secretly with his pupil. Still others laid it to telepathy; the N rays (a fad of the moment) and to " suggestion."
A scientific commission was appointed, which delivered its report on the 12th of September, 1904. It was com- posed of Messrs. Stumpf and Nagel, professors of psychol- ogy and physiology respectively at the University of Berlin, of the director of the Zoological Garden, a director of the circus, veterinarians, and cavalry officers. The report de- nied the existence of any tricks used by von Osten, for the very good reason that the experiments had the same results even in his absence. For the rest, the conclusions of the commission w^re entirely negative. It was said merely w^hat did not exist, but the answer to the puzzle was not revealed. The mystery deepened. Haeckel, who had as- sisted at one of the seances, declared for the authenticity of the mental activity of the horse. A second commission was formed, composed of Stumpf and his students. This time one point was seemingly established. Hans could not read, calculate, or count unless some one present knew the answers. Pfungst established this, indeed, beyond even a doubt by putting a series of questions to the horse, the answers to which were not known to the questioner. A card containing a certain number was picked at random by the experimenter and exhibited to the horse in such a way that no one could see it (the answer being unknown to the experimenter himself). When tested in this way, procedure without knowledge, the horse returned only Sfc of correct responses, whereas he responded with an ac- curacy of 9Sfo in those cases where the questioner knew the answers. Exactly the same state of affairs was found to exist with respect to arithmetical calculations and to reading. It will be recalled that reading was accomplished Jby first tapping out the number indicating the horizontal 302 THE LIMITS OF TRAINING IN ANIMALS row in which the first letter appeared and next the place in the row. This complicated procedure had to be repeated for each letter in the word. It w^as found that when the questioner was ignorant as to the letter and row, Hans never responded correctly. Calculation was tested simi- larly. It was found that if von Osten whispered a number in the horse's ear so that no one present could hear, and then the questioner similarly gave the horse a number and told it to add the two, the answer was wrong 28 times out of 31, whereas in a procedure with knowledge the relation was 29 right and 2 wrong. Hans could not even count on the abacus when the questioner remained in ignorance of the number of balls which were pushed aside. The memory tests and tests on musical ability (all questions of this kind being answered by tapping) all failed to yield correct responses in a procedure without knowledge. Taking all such tests into consideration showed that where the ques- tioner knew the answer 90-100^ of the responses were correct. "When the experimenter was ignorant of the answer, only lOfo at most of the answers were correct. These tests tell ns little as to how the experimenter 's knowl- edge aided the horse in giving correct responses. Carrying the analysis one step farther, Pfungst found that vision was necessary for correct response. When the horse was perfectly blinded almost no correct responses could be ob- tained (only 6^ were correct). The final link in unravel- ing the " mystery " came when Pfungst discovered the signs by means of which the horse responded. These sig- nals were the minimal movements of the questioner's head. ** As soon as the experimenter had given a problem to the horse, he involuntarily bent his head and trunk slightly forward and the horse would then put his right foot for- ward and begin to tap, without, however, returning it to its original position. As soon as the desired number of taps were given, the questioner would make a slight upward jerk of his head. Thereupon the horse would immediately swing his foot in a wide circle, bringing it back to the original position. (This movement which, in the follow- ing exposition we shall designate as ' ' the back step, ' ' was THE HORSES OF ELBERFELD 303 never included in the count.) Now after Hans had ceased tapping the questioner would raise his head and trunk to their normal position. This second, far coarser move- ment, was not the signal for the back step, but always followed it. But whenever this second movement was omitted, Hans, who had already brought back his foot to the original position and had thereby put it out of com- mission, as it were, would give one more tap with his left foot." These minimal movements turned out to be ex- ceedingly small (1/5 mm. and upward). If this analysis of Pfungst's is correct the responses of Hans are exactly similar to those with w^hich we have long been familiar in the laboratory.
The horses of Elberfeld. — Not every one accepted Pfungst's conclusions, yet interest in von Osten and in Hans rapidly abated. Von Osten died in 1909. Hans was entirely forgotten until Karl Krall's book appeared which raised the question of the completeness of Pfungst's explanation. Mr. Krall, a wealthy and highly respected merchant of Elberfeld, having known von Osten and hav- ing worked with him during his last years, inherited Hans. He states that Hans works well imder circumstances where it is impossible for him to see his master or to distinguish the minimal movements of the questioner. E.g., Krall affirms that Hans replies correctly in complete dark- ness, or with blinders on that permit him to see only the blackboard or the objects about which he is questioned. Krall determined to render justice to his old friend by proving the truth -of von Osten 's contentions. He bought two stallions, — Muhamed and Zarif, — and began to train them.