METHODS OF SCIENCE; NUMBERS 11 



ser, dog, mammal, vertebrate, animal, living 

 thing, object, — these are successive products of 

 the great process of abstraction. Often as we 

 proceed in the direction of greater abstraction 

 and idealization, we eliminate little by little the 

 empirical material from which the abstractions 

 were derived, but it seems probable that the 

 empirical is never wholly eliminated. So men 

 gather beet roots, and, subjecting the juices to 

 various arts of refining, obtain a sugar which 

 still contains some traces of its earthy origin. 

 Perhaps we know what we mean by a pure 

 chemical, but no one has ever seen one; and I 

 doubt if we shall ever see a pure abstraction. 



It would be far from my purpose to enter 

 upon an exhaustive study of the difficult prob- 

 lems in the theory of knowledge. As is remarked 

 by Edgar Allan Poe,^ himself a remarkable 

 thinker, "The mental features discoursed of as 

 the analytical, are, in themselves, but little 

 capable of analysis." Nevertheless, if we are to 

 discuss the concepts of science we must at least 

 have a partial realization of the mode of their 

 evolution. 



I say evolution rather than manufacture in 

 order to accentuate the point of view that I 



5 Edgar Allen Poe, Murders in the Rue Morgue. 



