THE REAL WORLD 371 



of Newtonian mechanics— but in this long period I do not see any 

 fundamental revision required, or made, in that system. Suppose 

 even that I have overlooked several such major revisions: are there 

 not obviously many successes of the theory any one of which out- 

 weighs a thousand failures? When Adams and Leverrier predict that 

 a new planet will be seen if a telescope is pointed at a certain seg- 

 ment of the sky at a certain time, what are the odds against their suc- 

 cess if the Newtonian system is but a clever correlative device? To 

 say the odds are astronomical is ambiguous but correct in every 

 sense. And to dismiss such developments as "chance," when they oc- 

 cur not once but repeatedly, is to masquerade a miracle in the guise 

 of chance. 



I maintain that theories selected for their heuristic power are 

 theories selected by an at least partially objective criterion. And the 

 demonstrable heuristic power of major scientific theories I hold to be 

 a genuine discovery demanding explanation. Heuristic power is a 

 fact; a fact unexplained and inexplicable within the phenomenalist 

 view of scientific theory— as nothing but a humanly invented system 

 for comprehensive correlation of economic descriptions of human 

 "sensations" experienced in the past. The obvious bankruptcy of this 

 position is, indeed, just what drove Duhem to his trans-phenomenal- 

 ist conception of scientific theory as a "natural classification." For 

 Duhem did not fail to recognize that the immense heuristic power of 

 a major scientific theory must signify that it offers some "reflection of 

 the real relations among the invisible realities." 



The highest test, therefore, of our holding a classification as a 

 natural one is to ask it to indicate in advance things which the future 

 alone will reveal. And when the experiment is made and confirms the 

 predictions obtained from our theory, we feel strengthened in our 

 conviction that the relations established by our reason among ab- 

 stract notions truly correspond to relations among things. 



Simplicity. On ever-narrower postulational foundations, we en- 

 deavor to construct scientific theories of ever-increasing scope and 

 precision of correlation. Human desire explains why we seek such 

 theories but, as Weyl observes, not why we are able to find them. 



. . . the farther the analysis progresses, the more detailed the ob- 

 servations become and the finer the elements into which we dissect 

 the phenomena, the simpler— and not the more complicated, as might 



