236 MIND IN EVOLUTION CHAP. 



the experiments suggest " practical ideas " negatively, 

 they strongly suggest absence of any sort of analysis in 

 the genesis of those ideas. An animal can shift its 

 attention to this or that object or change within the 

 sphere of perception ; but it cannot apparently follow out 

 the structure of any complex object with any minuteness 

 and accuracy. 



8. Further evidence as to perceptual learning. 



To return to perceptual learning. If we allow ourselves 

 to go outside the region of experiment we find plenty of 

 instances of it in the observations of psychologists and 

 others. Of these the ordinary case of the cat and the 

 thumb-latch will serve, as an example, as well as another. 

 Writers like Mr. Thorndike, who deny imitation, have to 

 fall back on the theory that the cat learns this, in the first 

 instance, by some lucky combination of random jumpings 

 and clutchings and scratchings and pushings. This would 

 only be possible, to begin with, if cats habitually jumped 

 about and clutched and scratched in such a manner when 

 they wanted to leave a room. Mr. Thorndike has experi- 

 mented so much with cats shut up in cages that he seems 

 to think this a possible explanation. If three cats out of 

 eight learnt, by scrabbling all round their pens, to open 

 their latches, why, he asks, 1 should not three cats in a 

 thousand learn to open a thumb-latch "in the same way"? 

 i.e., I suppose, by jumping and scratching all round the 

 room. But the reply is that a thousand cats do not 

 spend their time jumping and scratching round rooms. 

 Such behaviour is, no doubt, natural to cats confined in a 

 cage, but ordinary observation of cats under normal 

 conditions is enough to show that this explanation is 

 purely imaginary. 



Comparative unimportance of accident. 



I must add here that, so far as I could judge, the more 

 a success was accidental the less likely were the animals 



That these elements may bleach out and attenuate into ideas is not 

 impossible" (pp. cit., p. 155.) 



As I use it, the term idea means any mental state, however little 

 analysed, the function of which is to refer to something not actually 

 perceived (see above, Ch. VIII., p. 151). 



1 Op. tit., p. 44. 



