March 31, 1911] 



SCIENCE 



All 



high stage of organization has been at- 

 tained. The evolution along the lines of 

 complexity of instinct and ready modifi- 

 ability of reactions to suit new conditions, 

 affords a substantial basis for intelligent 

 behavior. Without such evolution the 

 power of associative memory would avail 

 little. But with a large number of readily 

 modifiable instincts, associative memory be- 

 comes the means of affording a much wider 

 and closer adjustment to the environment. 



The studies which have been made of 

 primitive types of intelligence such as 

 found in crustaceans, fishes and amphibi- 

 ans have shown that associations are 

 formed by a gradual process of reinforce- 

 ment or inhibition of a particular reaction 

 to a given stimulus. The method followed 

 is one which Lloyd Morgan has designated 

 as "trial and error." It may be illus- 

 trated by the experiment of Terkes on the 

 formation of associations in the crayfish. 

 In these experiments a box was employed 

 into one end of which the crajnfish was 

 admitted through a narrow aperture. The 

 other end of the box was divided by a 

 median partition which gave the crayfish a 

 choice of two routes to a tank of water at 

 the other end into which the creature was 

 naturally desirous of getting. One of the 

 two ways to the water was closed by a glass 

 plate at its farther end so that the crayfish 

 was afforded a choice of a right and a 

 wrong path to the water. Would the cray- 

 fish after a number of trials learn to choose 

 the right path and avoid the closed pas- 

 sage? In the first ten experiments the 

 crayfish went as often to the right as it did 

 to the leftj but in the next ten trials the 

 percentage of correct choices was some- 

 what greater. Finally after a large num- 

 ber of trials the animal came to choose the 

 right path to the water, making but rarely 

 any mistakes. 



Similar experiments with crabs, fishes 



and the frog have yielded similar indica- 

 tions of slow learning. In some respects 

 such learning resembles the slow formation 

 of a habit rather than the judgment of a 

 consciousness which "sizes up" the situa- 

 tion and determines upon a certain course 

 of action. It is quite probable that such a 

 primitive form of learning does not in- 

 clude any association of ideas. It can be 

 satisfactorily accounted for by assuming 

 nothing more than an association of cer- 

 tain sense perceptions with particular 

 movements. The animal may have no 

 ideas to associate — nothing but sense im- 

 pressions and motor impulses. Of course 

 its mental content may include much more 

 than this, but in interpreting the behavior 

 of animals it is generally advantageous to 

 follow the principle laid down by Lloyd 

 Morgan — which is a sort of special appli- 

 cation of the law of parsimony — that we 

 should not assume the existence of a higher 

 psychic function if the phenomena can be 

 explained as well in terms of a lower one. 



The step from sensori-motor association 

 to the association of ideas is not, I believe, 

 a wide one, and comes about as a natural 

 consequence of the elaborateness and what 

 Hobhouse has designated as the "articu- 

 lateness" of the mental process of adjust- 

 ment. It is foreign to our purpose, how- 

 ever, to trace the increase in the number, 

 delicacy, quickness and complexity of the 

 processes of association which we meet in 

 the various stages of mental evolution. 

 One problem at present lies in the initial 

 step involved in the formation of a simple 

 association. And it is a problem which, 

 despite its apparent simplicity, involves 

 the consideration of some vexed and subUe 

 questions. 



In learning we have to do with two op- 

 posite processes of reinforcement and in- 

 hibition. A chick after it pecks at a cater- 

 pillar which is wholesome and savory pecks 



