TRIAL AND ERROR REACTIONS IN MAMMALS 59 



subjects who were able to elaborate a succession of previous 

 experiences in such a manner as to associate, at each trial, the 

 door last opened with an awareness equivalent to " during any 

 present trial the door last opened is apt to be impossible as a 

 means of exit." In other words, there would have to be an 

 association in which one of the elements would be a complexly 

 derived awareness of a principle deterrent to the activity in- 

 volved in trying the unlocked door of the immediately preceding 

 trial. The conscious avoidance of the impossible door as such 

 may therefore be looked upon as due to a tendency to make 

 rational inferences from a sequence of past experiences. For 

 convenience of statement it will be referred to hereafter as 

 " the rational inference tendency." 



Type B. This reaction involves trying all four doors, but 

 once each, and in an irregular order. 



From the normal adult human viewpoint this adjustment 

 contains but one error, viz., the inferentially and actually use- 

 less effort to open the impossible door. In cases where there 

 is clearly apparent a perception of the impossible door principle 

 (i. e., where there is given a record of appreciably more than 

 50 per cent of Type A reactions), the manifestation of Type B 

 reactions may be assigned either to mere inattention or to lapse 

 of memory, or to the necessity of falling into the impossible 

 door error a certain number of times before an awareness of 

 the principle can be obtained. If we exclude these cases, the 

 Type B reaction may be looked upon as an expression of a high 

 type of searching tendency. Although it shows a lack of modifi- 

 cation by the higher tendency to make rational inferences, it is 

 of much significance as showing the absence of modification by 

 the low^er tendencies to which we shall ascribe reactions of the 

 C, D and E types. Of course the inclusion of an effort to open 

 the impossible door may be due, in a certain number of cases, 

 to the interference of a tendency to associate the last successful 

 activity with the impossible door, in which case the searching 

 tendency is not the sole reactive factor. But since we are deal- 

 ing only with the reactive tendencies that precede the establish- 

 ment of definite and habitual associations, we may, for con- 

 venience, refer to Type B reactions as due to " the unmodified 

 searching tendency." This, of course, only when the number 

 of Type A reactions is so small as to exclude the possibility that 



