34 THE LIMITATIONS OF SCIENCE 



double refraction and polarization of light were dis- 

 covered by Huygens before he, and later Fresnel, 

 attempted to explain them as a modification of me- 

 chanical waves in an elastic solid ether; and Newton 

 announced the laws of the interference of light which 

 occurs when there are reflections between thin films 

 before he pictured this phenomenon by his hypothesis 

 of corpuscles. To-day all these laws remain while the 

 specific hypotheses have been discarded. How can we 

 say the hypothesis of atoms and ether led to dis- 

 covery in these cases? The fact is just the contrary: 

 hypothesis, at least that part of it which consists in 

 developing a mechanical model of the action, follows 

 experimental discovery; it is the effort to explain or 

 visualize the unknowable processes involved in known 

 experimental facts and mathematical laws. 



In the second place, these metaphysical hypotheses 

 progress from the simple to the complex. Each new 

 fact discovered adds its quota to the irreconcilable and 

 conflicting properties of the ether and the atom, and 

 these invisible links of the universal machine grow 

 more and more bewildering and complicated, until the 

 whole construction falls to pieces. Nor is this all ; the 

 man of science forgets that he is building toy houses, 

 and ends by believing in their reality. Even if hypoth- 

 esis does not carry him so far, it certainly has this 

 effect on others who accept the dogmas of science 

 without discrimination. It is no small danger thus to 



