592 KARL T. WAUGH 



When Thorndike put his cats into a cage, the process of learning 

 to open the latch consisted in the gradual association of a certain 

 movement with certain sense impressions under certain condi- 

 tions of hunger. Under the instinctive excitement caused by the 

 situation the cat makes many movements; those in each part of 

 the cage are guided by the visual impressions of that part of the 

 cage acting by way of the visual cortex. Each group of visual 

 impressions would thereafter, in accordance with the law of neural 

 habit, tend to lead to the same movements more readily than 

 before. One of these impressions acquires increased intimacy of 

 association with a certain movement more readily and certainly 

 than the rest, viz., the visual impression made by that part of the 

 cage in which the latch is situated, with that movement which 

 results in the falling open of the door. The intimacy of this par- 

 ticular association is increased each time this particular movement 

 is made, until, as the cat casts its eye over the cage, the visual 

 impression of that part of the cage at once evokes the movement. 

 The explanation which Thorndike gives as to why the association 

 between the particular sensory path and the particular motor 

 disposition becomes fixed while other possible motor dispositions 

 do not become fixed, is, that the former association gets " stamped 

 in" by the pleasure resulting from it, while the other is "stamped 

 out" by the pain of failure. 



The kinaesthetic sense undoubtedly has a tendency to deter- 

 mine the cat's behavior, but vision operates more quickly, for the 

 cat directs its attention to the visual stimuli, rendering possible 

 readier association between the object seen and the motor mechan- 

 ism. The condition here differs only in degree, not in kind, from 

 that obtaining in the mouse. 



The relative importance of muscular association and visual 

 association may be well shown by the analysis of the actions of a 

 mouse used in problem 2, B. If we go on the assumption that the 

 mouse's action was associated with the result of the action immedi- 

 ately preceding, then we divide the whole series of choices of the 

 animal into two kinds of sequences : position sequences, in which 

 the mouse turns to the same side, left or right, where it received 

 food in the preceding trial, and avoids the side where it received 



