CHARACTER OF SPECULATIVE PROBLEMS 397 



"reality (loud cheers)." 1 Many of the conclusions of R. A. 

 Millikan with reference to problems of religion, though 

 they are hardly to be called wishful thinking, are so little 

 based upon the data of science and so significantly founded 

 on a deep, personal religious experience, that they are hardly 

 to be classed as attempts at speculative philosophy in the 

 narrow sense in which the term is here being used. The 

 scientist who becomes enraptured by the beauty of a deduc- 

 tive scheme, who marvels at the grandeur of the heavens, 

 who wonders at the complexity of life — and concludes 

 therefrom a Spirit behind nature, is not making a judg- 

 ment of fact based upon the data of science but a judgment 

 of value based upon an esthetic experience. But if it is the 

 emotional man who is speaking, is one justified in using 

 the abstract, rational man as the judge of what is being said? 



It seems important to recognize, therefore, in the fourth 

 place, that speculative problems almost always involve 

 implicit recognition of data which are other than those 

 offered by science. Eddington cannot conclude anything 

 as to the existence of a Spirit behind nature from the facts 

 of science alone, for scientific symbols are absolutely incap- 

 able, on his own grounds, of providing such. information; 

 it is rather the mystic experience and the introspective 

 experience which justify the inference. Bergson concludes 

 to the mobile and enduring character of reality not through 

 science, for science can reveal only the static and discon- 

 tinuous, but through intuition. Compton talks about 

 problems of religion not as a scientist but as a man attempt- 

 ing to guide his own individual life. 2 It often turns out, 

 therefore, that the essential inference to the extra-scientific 

 domain is based upon data which are also extra-scientific, 

 and the speculative problem then reduces to one of showing 

 that the existence of such a domain is compatible with the 

 data of science. The inference is, consequently, not from 

 science to philosophy, but entirely within philosophy (or, 

 at least, entirely outside of science). The result is that the 



1 Ibid., p. 287. 2 The Freedom of Man, pp. ix-x. 



