TENDENCIES OF MODERN PHYSICS 37 



sweep aside so large a part of scientific thought as 

 hypothesis has been, unless it is really parasitical, a 

 hindrance rather than an aid to development. This 

 opinion as to the uselessness of hypothesis seems to be 

 gaining ground. Thus M. Duhem, in his Theorie 

 Physique, states that physical theories must have one 

 of two aims: either to explain laws which have been 

 established from experience, or to classify such laws 

 without giving any explanation. Of the two, the sec- 

 ond only is a legitimate scientific process, as the first 

 method makes physics dependent on metaphysics and 

 so introduces occult and unverifiable causes. A proper 

 theory should thus give us a classification of laws 

 and should point to new experimental methods, thereby 

 tending to intellectual economy in that we are per- 

 mitted to forget a multitude of details and otherwise 

 isolated facts in one common expression. He further 

 claims that the construction of a mechanical model 

 as an explanation of a law does not lead to such dis- 

 coveries, since these are really derived from abstract 

 principles, the model being invented afterwards merely 

 to make the law concrete. In this opinion he is sup- 

 ported by Hertz, who, after discovering experimentally 

 the electric waves predicted by Maxwell, found the best 

 statement in Maxwell's equations, and not in his model 

 of ethereal lines of force. 



A ruthless and complete elimination of hypothesis is 

 undoubtedly impossible, and is not even desirable. We 



