CHARACTER OF SPECULATIVE PROBLEMS 393 



may mean either that man is free in the sense that his actions 

 do not violate any of the laws of physics (Eddington and 

 Compton) or that he is subject to causal laws but unpre- 

 dictable in his behavior (Planck). Both views permit man 

 to do as he pleases, and this seems to be the demand of 

 human freedom. 



(3) The nature of reality, (a) The data in this problem 

 are of the greatest variety. For three of the writers who will 

 be considered, the data are features of method; for the 

 fourth, they are a feature partly of subject matter and 

 partly of method. The most common solution to this prob- 

 lem takes its origin in the recognition of the essentially 

 limited character of science; the scientific method is shown 

 to predetermine the sort of object which the scientist will 

 find and to prevent him from finding an object of another 

 kind; for example, because the scientist employs mathe- 

 matical techniques he discovers only mathematical objects, 

 and because he excludes preferential judgments he fails to 

 find values. Accordingly, there is almost always drawn into 

 this problem, as an essential datum, the fact of an inde- 

 pendent knowledge of the non-scientific objects. Though 

 the scientist does not find God in his laboratory he does find 

 him when he leaves the laboratory and begins to live his 

 life in the broader sense of the term. Hence the problem in 

 this case often consists in showing the compatibility of the 

 conclusions of science with the convictions derived through 

 intuition, mystic insight, and revelation, (b) The inferences 

 are correspondingly various. In three of the cases to be 

 examined they proceed from the character of scientific sym- 

 bols to the character of that which is symbolized; but in 

 two of these cases the referent of the symbol is considered 

 to be essentially unlike its symbol, and in the third case 

 essentially like its symbol. In all three cases the inference 

 has some plausibility only because it is based upon an inade- 

 quate theory of the nature of symbols. In the fourth case 

 the inference is analogical, and passes from series which 

 have limits to other series which appear to have no limits 



