c 



Essays in Statistics. 183 



finally they attain to a close relation or identity, the most perfect 

 end which they can reach. In this process of evolution there are, 

 as it seems to us, two principal stages : the first of these, which 

 constitutes science properly so called, consists in the employment 

 of verification; the second constitutes exact and ideal science, 

 and it consists in quantification, or, to avoid neologism, quantita- 

 tive determination. 



This we will try to show. 



When we are aware of a large number of phenomena which are 

 analogous, that is, at once like and unlike, we endeavour to seize 

 the fixed basis in the production of these phenomena their law. 

 But whether this law result from an intuition of genius, or from a 

 slow and minute comparison of facts, followed by induction, must 

 be submitted to the process of verification, for it has to explain all 

 the facts, or at least most of them ; and it alone must explain them, 

 otherwise it remains an hypothesis. 



Thus every science, in order to become science, passes through 

 three stages, the facts, the law, and the verification. First, the 

 phenomena are collected and observed, scrutinized, turned over and 

 over, placed on the rack of experiment, then from them is drawn 

 their generic constant element ; finally, the law thus discovered is 

 anew tested by application to facts, just as a seal is verified when 

 applied to its impression. This last process verification is 

 essential. 



. Without verification there is no science, because this process 

 alone can give to our theories an objective value. It is a 

 complete mistake to suppose that what is not true can be 

 scientifically established. There are a hundred ways of looking 

 at facts, of interpreting, and of generalizing them. Of course, these 

 are not all correct, but who is to decide between them ? In such 

 case science gets only the individual, personal opinion of one man, 

 his special mode of understanding and accounting for the facts. 

 But this is an entirely subjective doctrine, which may indeed be 

 science, but if so is science only by accident, nor have we any 

 means of knowing that it is science, or any grounds for affirming 

 that it is. 



It may be said, parenthetically, that this is what distinguishes 

 metaphysics from science. 

 9 



