48 DOGMATISM AND EVOLUTION 



uncrossed gap still remains between the truths of reason and the 

 generalizations based upon experiment. 



Can we go further and say that the rationalistic deduction is 

 not only indefinitely far from accomplishment, but essentially 

 impossible? Perhaps not; for the simple reason, that whenever 

 a law is shown to be indemonstrable from an existing set of axioms, 

 the rationalist may be expected simply to claim it as an additional 

 axiom. At the same time, the prospect of such procedure on his 

 part has a fatal effect upon the convincing power of his system. 

 Axioms are supposed to be derived, not from the exigencies of 

 the demonstrations that are founded upon them, but from the 

 unbiased intuition of reason. A defect in the system is thus 

 disclosed which is analogous to the initial difficulty involved in 

 the assumption of synthetic axioms that even if it assumed a 

 form in which (in the existing state of the sciences) it was wholly 

 irrefutable, it would remain powerless to convince a sceptic. 



We have been led so far in the discussion of the rationalistic 

 side of our subject that we must confine ourselves to a few words 

 upon its empiricistic side. However, a very few will suffice. 

 That the doctrine of the externality of relations is involved, for 

 empiricism as well as for rationalism, in the postulate of simple 

 elements has already been shown. That, furthermore, these 

 presuppositions lead the empiricist unavoidably to an atomistic 

 chaos in which all relations disappear, is practically admitted 

 by Hume in the remarkable appendix to his Treatise, and it has 

 been joyously reaffirmed by his critics ever since. We have seen 

 that the rationalist commonly saves himself from a similar embar 

 rassment by the assumption of synthetic axioms. It is worth 

 while inquiring how the empiricist avoids, or at least postpones, 

 the fatal reductio ad absurdum. Berkeley and Hume are here, 

 in sharply contrasted ways, representative of their school. Each, 

 it will be seen, calls in rationalistic principles to his aid; 1 and 



^his has already been shown for Hume in the course of our discussion of 

 immediate experience. 



