148 EVOLUTION AND MAN S PLACE IN NATURE 



for intelligence in animal life, 1 I shall best meet 

 the requirements of the present argument by giving 

 prominence to definite generalisations from the data 

 supplied. The evidence has been largely accumulated, 

 and is now generally accessible in literature. There is, 

 therefore, good reason for anticipating a solid basis of 

 agreement here, even though the difficulties of the in 

 quiry are such as to preclude the hope of an exhaustive 

 discussion. We shall never wholly escape the per 

 plexities arising from difference between the indirect 

 mode of knowing how animals act, and the direct 

 knowledge of our own intellectual procedure. External 

 observation has no chance of competing on this field 

 with consciousness of individual experience. Nothing 

 but defective appreciation of the difference in the 

 conditions of knowing, can induce any one to suppose 

 that it is possible to reach conclusions as to animal 

 intelligence with the certainty attainable as to human 

 intelligence. We know beyond doubt what are the 

 characteristics of rational power in man; we can 

 never know, except by very imperfect inductions, 

 what powers of intelligence are at work in animal life, 

 There is, moreover, this additional obstacle to exact 

 inference, that the certainty as to our own procedure, 

 over against the uncertainty as to animal procedure, 

 leads us to reason largely by analogy, without our 

 having any strict test of the legitimacy of the 

 analogical inference. All these limitations and dis 

 advantages must be kept in view, if scientific con- 

 elusions are to be sought with reasonable hope of 

 success. Full weight must, therefore, be given to 



i The Relations of Mind and Brain, chap. vii. pp. 198-288. 

 Revised treatment of the question appears in the 3rd edition. 



