THE EVOLUTION OF SCIENTIFIC THEORIES 269 



mentally new theory will, on first presentation, almost always be 

 flatly rejected? However, this is no reductio ad absurdum but, rather, 

 just what Taton observes to be historically the case. 



Every great discovery has produced some sort of intellectual scandal, 

 has been opposed by current, and always badly informed, opinions on 

 the basic nature of scientific problems, and also by the majority of 

 scientists of the time holding outdated theories, and incapable of re- 

 nouncing some of their most solidly ingrained ideas. 



The first two criteria taken singly, or together in appraisals of 

 "simplicity," are then insufficient to ensure sound judgment of new 

 theories. Presumably we must pin our hope for such judgment to the 

 third criterion: heuristic power. Yet the first two criteria are not then 

 inconsequential. Before it can display its heuristic power, a new 

 theory must win at least provisional accreditation by men willing to 

 accept its guidance in their work. Such accreditation can derive only 

 from the sense that the new theory offers some promise, if not yet any 

 realization, of superior correlative efficiency and/or explanatory ap- 

 peal. Consider too that, as earlier noted, both the correlative and 

 explanatory functions enter into the heuristic function. We saw in 

 the preceding chapter how prominently the explanation offered by a 

 theory figures in its use as an instrument of discovery. And of course 

 the theoretic correlation permits us effectively to marshal the knowl- 

 edge we already have in the quest for new knowledge. Finally, as 

 we will see, the heuristic functioning of a theory may so alter the 

 balance of efficiency and appeal that in the long term "simplicity" ac- 

 quires a large significance it lacks in the short. 



HEURISTIC POWER 



Duhem minimizes the explanatory function of a scientific theory, 

 stresses its correlative function, and never renders account of its 

 heuristic function. Dirac, too, minimizes the explanatory function, 

 but only to emphasize the supreme importance of the heuristic 

 function. 



The only object of theoretical physics is to calculate results that can 

 be compared with experiment, . . . 



. . . The object of quantum mechanics is to extend the domain of 

 questions that can be answered and not to give more detailed answers 

 than can be experimentally verified. 



