58 G. V. HAMILTON 



(5) The relation of every present trial to its immediate pre- 

 decessor was such as to exact a penalty of non-success for trying 

 the unlocked door of the immediately preceding trial. 



(6) During a given trial, a second or third, etc., effort to 

 open a particular door was invariably unsuccessful: a door 

 which would not yield to a single, definitely directed attack 

 during a given trial could not possibly be opened by the sub- 

 ject during that trial. 



(7) The various doors were discriminable, one from another. 

 Under the conditions just enumerated, the various types of 



reaction that we have isolated are capable of the following 

 psychological interpretations : 



Type A. To conform to this type, the reaction must include 

 a single, definite effort to open each of the three inferentially 

 possible doors, and must not include an effort to open the infer- 

 entially impossible door. It will be remembered that the impos- 

 sible door varied from trial to trial. 



Now it is evident that only an awareness of the impossible 

 door as such would enable any subject to manifest appreciably 

 more than 50 per cent of Type A reactions out of his total number 

 of classified reactions. If the impossible door were of indifferent 

 value for reaction, either as impossible or as the object of the 

 subject's latest successful activity, and if during no trial the 

 subject were to make more than a single effort to open any one 

 door, his record would tend to show 50 per cent of Type A reac- 

 tions and 50 per cent of Type B or C reactions. Of course, a 

 preference for the impossible door or a tendency to make more 

 than a single continuous effort to open a particular door during 

 a given trial would impair the subject's chance of approximating 

 even 50 per cent of Type A reactions. 



What reactive tendency, then, would lead to more than 50 

 per cent of Type A reactions? It is obvious that this cannot 

 be the primitive tendency to reduce diffuse activity-impulses to 

 a definite attack upon a single object. The situation rendered 

 it impossible for any subject to associate a simple object with 

 successful activity and to obtain thereby a formula for invariable 

 success. On the other hand, the establishment of a simple nega- 

 tive association was not sufficient to enable any subject to avoid 

 the impossible door. This spatially varying object-to-be-avoided 

 could clearly obtain its true value for reaction only among the 



