628 



SCIENCE 



[N. S. Vol. XXXIX. No. 1009 



years. Then, observation was naturalistic, 

 general: now, it is experimental, specific. 

 Most obvious and most important of the 

 changes is the development of methods by 

 which the conditions of observation may be 

 controlled and the results precisely re- 

 corded. The older work was, on the whole, 

 crude, inconclusive, wasteful of time, lead- 

 ing to few safe generalizations ; the newer, 

 by contrast, is precise, carefully controlled, 

 and tending to lead to the formulation of 

 laws. It is, of course, desirable that we 

 recognize that not all observation can be 

 conducted under experimental conditions, 

 that the naturalistic method in the study of 

 human behavior as also in the study of 

 animal behavior has its proper and impor- 

 tant place. "We should gladly recognize its 

 values, while insisting upon the importance 

 of supplementing it by the experimental 

 study of the same phenomena. 



But the practical-minded person has 

 doubtless been asking, throughout this in- 

 troductory discussion, of our topic, "How 

 may the reactions of a person be scientif- 

 ically studied ? Is it possible, ordinarily, to 

 subject a human being to such conditions 

 of observation as are used in experiments 

 with other animals?" A few examples 

 from studies of human and infra-human 

 behavior will serve as an answer to these 

 questions. 



One of the most interesting aspects of 

 organic activity is its modifiability. We 

 designate this as habit-formation. Now, it 

 happens that in a great variety of organ- 

 isms the formation of habits has been 

 studied- experimentally. In the case of the 

 dancing mouse,^ for example, I investigated 

 the relation of rapidity of habit-formation 

 to certain external conditions. With an 



* Yerkes, Robert M. and Dodson, John D., "The 

 Relation of Strength of Stimulus to Rapidity of 

 Habit-formation," Journal of Comparative Neur- 

 ology and Psychology, 1908, Vol. 18, p. 459. 



apparatus so arranged that the mouse could 

 choose as its route through the experiment 

 box either a dark or a light passageway, I 

 determined the number of experiences 

 necessary in order that the animal should 

 learn that, no matter what the spacial rela- 

 tions of the passageways, only the light one 

 could safely be chosen as a way of egress, 

 since each time the dark passageway was 

 entered a disagreeable electric shock re- 

 sulted. It was the problem of the observer 

 to discover how quickly, under a given set 

 of conditions, the mouse would learn always 

 to avoid the one passageway and to seek 

 the other. In this experiment, which was 

 so conducted that strictly comparable re- 

 sults were obtained from several individ- 

 uals, it was first ascertained that the less 

 the difference in lightness of the two pas- 

 sageways, the longer it took the mouse to 

 learn to choose correctly. Next, it was 

 determined that the rapidity of learning 

 varied with the strength of the electric 

 shock, which was regularly given as punish- 

 ment for mistakes. When the passageways 

 differed markedly (discrimination easy), 

 the stronger the shock the more rapid the 

 learning. When the passageways differed 

 slightly (discrimination difficult), beyond 

 a certain point, increase in the strength of 

 the shock delayed the learning process. 

 When the passageways differed by an inter- 

 mediate amount, it appeared that an inter- 

 mediate strength of stimulus was most 

 favorable to habit-formation. From these 

 observations, it was possible to deduce the 

 following law for the behavior of the danc- 

 ing mouse: As the difficultness of visual 

 discrimination increases, that strength of 

 electric stimulus which is most favorable as 

 a condition for the acquisition of a habit 

 tends to approach the threshold. 



In this investigation, we have, first, a 

 definite problem ; second, a reasonably large 

 number of observational data (facts of be- 



