STUDIES ON THE MENTAL LIFE OF ANIMALS. 565 



It also appears to us that Mr. Thorndike's definition of imitation is 

 rather a narrow one; starting from this definition he concludes that 

 imitation does not exist among animals. According to him, in order 

 that there should l)o imitation the subject must exactly reproduce the 

 same movement, the same gesture that he has seen. So (see text) cat 

 No. T did not imitate cat No. 1, because the latter opened the door by 

 pulling a loop with its teeth, while No. 7 pulled it with its paw. It 

 seems to us, however, that the exact, complete reproduction of a 

 movement is an example of imitation that is I'elatively difficult and 

 quite complex: it is the perfect form of imitation, very common with 

 man, who possesses a developed imagination. There may be some 

 exaggeration in expecting to find it as perfect as this among animals. 

 The subjects of Mr. Thorndike have not reproduced the movements 

 of their model because they did not possess a sufficiently strong or 

 lively faculty of representation. The experiments cited by us may 

 likewise be interpreted otherwise than is done by the author. The 

 point was to see whether the animal, guided by the author, would 

 repeat the movement which the latter made him execute. It was 

 required to open the box C, closed by a button. It was shown that 

 the subjects did not turn the button in the way they had been taught; 

 they did not repeat identically the act wdiich it was desired to teach 

 them. But the essential matter is that there was formed in their con- 

 sciousness the notion, even though vague, of a relation between the 

 situation before them and the act of turning the button. Each subject 

 would then translate this relation in a manner personal to himself — 

 one would push the button with his nose; another would bite it, etc. 

 In brief, we would not have an exact reproduction of the movements 

 which had been taught — there would not be a perfect imitation; but 

 the question remains unaffected, whether the primai-y. rudimentary 

 form of imitation does not consist, in the experiments here cited, 

 merely in a connection between a given situation and a possible direc- 

 tion of movements. The button of the door does not appear to the 

 subject as the central point on which the action depends; progress in 

 imitation will consist in associating with the point the image of the 

 special movements executed by the animal that served as model. We 

 must expect to find in them a trace of imitation, a tendency still vague 

 in its manifestations, rather than a, clearly established habit. The 

 disadvantage of Mr. Thorndike's method appears to us to be that he 

 has transported into the domain of animal psychology the notion of 

 imitation defined according to human examples. Unwittingly, Mr. 

 Thorndike has not escaped the prejudice that consists in observing 

 animals with preconceptions derived from the normal psychology of 

 man. In this very question of imitation the problem appears to us to 

 be that of seeking within the animal series to find what acts, what 



