CHARACTER OF SPECULATIVE PROBLEMS 395 



such, problems of this type are distinguished from questions 

 of pure method, and from questions of the meanings of the 

 basic concepts, for in both of these latter cases there is 

 reason to believe that the philosophical problems are in 

 some very intimate way tied up with the scientific problems, 

 and hence to be looked upon as their logical development. 

 A good scientist must be a logician and a critical philosopher. 

 But a good scientist need not be a speculative philosopher. 

 Many, in fact, would insist that a scientist who pursues 

 speculative enterprises ceases thereby to be a scientist, for 

 he abandons critical methods. This may be true, but at 

 least the scientist is under no obligation to occupy himself 

 with these problems so far as the strict demands of his 

 science are concerned. It is still true, as Laplace remarked 

 to Napoleon, that the astronomer does not need the hypoth- 

 esis of God. The chemist, as chemist, need not concern 

 himself with the possible social applications of poison gas 

 and high explosives. Nor should the physicist refrain from 

 his investigations until he has succeeded in proving ab- 

 stractly that he is free to do as he pleases, or that there 

 really is a world outside of himself which is there to be 

 known. Science goes on its happy way independently of 

 these more speculative considerations. They belong, as 

 many would maintain, rather to the frills of science, to be 

 pursued as an entertaining and harmless pastime by those 

 whose minds turn readily to such speculative enterprises. 

 In the second place, speculative problems seem to depend 

 for their solutions upon the more critical problems of the 

 philosophy of science. Speculative philosophy, in other 

 words, presupposes either the logic or the metaphysics of 

 science. This seems to be due to the fact that the more 

 synoptic problems are founded, properly speaking, not 

 upon science itself but upon certain general features of its 

 method or certain general properties of its concepts. These 

 more philosophical aspects of science must be recognized 

 before the step to speculative considerations is possible. 

 For example, a classification of the sciences is based either 



