ix LEARNING AMONG HIGHER ANIMALS 183 



5. Failure to imitate. 



A more serious negative result was reached by Mr. 

 Thorndike on another point. He found reason to think 

 that imitation in the sense of "transferred association" does 

 not exist among mammals below the primates. That is to 

 say, that animals never will do a thing because they see 

 others obtain their ends by so doing. Cats and dogs were 

 shut up in boxes, and allowed to watch the method by 

 which another cat or dog escaped from another compart- 

 ment of the box. To this compartment they were 

 afterwards admitted, and it was found that they had learnt 

 nothing. These experiments are of somewhat doubtful 

 value. 1 So far as they go, they fall in with the view to 

 which I have been led independently, that an animal 

 makes no intellectual provision for the future. If the 

 cat could say to itself, " There is my friend Tom in the 

 outer compartment. I shall be let in there in a minute, 

 so I will watch, and see how Tom gets out;" it would 

 doubtless watch Tom's proceedings with care, but it would 

 also be a much more reflective animal than it is. The 

 probability is rather that its attention is aroused only as a 

 rule by the act or object that leads up to or is connected 

 with the end to which it is itself striving. Other things 

 may strike it by chance, but its interest is not directed to 

 them. Thus it would be more likely that one animal 

 would imitate another if both were trying to escape or to 

 get fed at the same time. 2 Whether under these conditions 



" This method manifestly disregards the factor of attention above adverted 

 to, and consequently fails to demonstrate the absence in animals of those 

 higher associative processes involved in inferential imitation " (p. 164). 

 It follows that the method would be much less likely to succeed with a 

 cat, which resents being held, than with the obedient elephant or the 

 trustful dog. I failed in my attempt to use it with a Rhesus monkey 

 (see below, p. 284). 



1 Mr. Thorndike says that the cats watched their instructors " more or 

 less," and afterwards reckons up the number of times which a cat 

 " probably saw " or " certainly saw." My own experience in experiment- 

 ing convinces me so strongly of the difficulty of judging what a cat or 

 dog really sees even when I am making a direct effort to get its atten- 

 tion that I am surprised that so experienced an investigator should 

 speak of certainty in connection with the question whether one cat saw 

 another perform a seemingly trivial operation on the other side of a 

 grating. 



2 Mr. Thorndike tried this experiment with two cats, of which, one, in 

 fact, succeeded. Mr. Thorndike, however, states that it acted just like 



