SCIENTIFIC DISCOVERY 185 



"fictions " since they create entities such as classes, series, and 

 complexes, not certainly known to exist. The question as to 

 whether or not operations of construction eventuate in fictions 

 is unimportant here. What is to be emphasized is merely the 

 kinship of classifying, ordering, and correlating, and the fact 

 that they are all, in a sense, more than mere descriptions. 

 For this reason it will be well to speak of the method of con- 

 struction in preference to the method of abstraction. 



But in the second place, the two methods should not be 

 looked upon as alternative, for neither can be employed 

 without the other, and taken together they constitute only 

 one method which is the method of science. This can be 

 made clear by examining the distinction between them. It is 

 the acknowledged aim of the method of construction to avoid 

 any unjustified addition to the data; the method merely 

 elaborates or clarifies by emphasizing features of the given. 

 The method of hypothesis, on the other hand, consciously 

 adds to the given "by inventing a fictitious substance or 

 process or idea," as Dingle says. Hence it seems fair to say 

 that the aim of the method of construction is to produce 

 nothing new, while the aim of the method of hypothesis is to 

 produce something new. Expressing the same idea in other 

 words, one may say that the aim of the method of construc- 

 tion is to make as precise as possible the data which are 

 relevant to the problem, while the aim of the method of 

 hypothesis is to devise a symbol by which these data may be 

 explained. This suggests that the method of construction 

 is not alternative to the method of hypothesis but prepara- 

 tory to it; it also suggests that the method of hypothesis is 

 not alternative to the method of construction but its neces- 

 sary outcome. Together, the methods make up the act of 

 discovery; singly they constitute the stage in which the 

 imagination is prepared, and the stage in which it makes its 

 leap into the unknown. 



This analysis permits a reinterpretation of the methods. 

 If one speaks of both methods as operations for the creation 

 of symbols he may then distinguish them in the following 



