THEORIES AND MODELS 245 



by associating it with an explicit particle model. Thus, for example, 

 Planck comes to his celebrated interpretation of blackbody radiation 

 through this species of thermodynamics. 



Is thermodynamics perhaps simply a poor exemplar of the great 

 heuristic power to be found in other theories with the highly formal 

 construction recommended by Duhem? Newtonian mechanics has 

 displayed immense heuristic power, and can be given a thoroughly 

 formal development, but we customarily envision perfectly concrete 

 models whenever we apply the formalism. Classical electrodynamics, 

 in the severely abstract form given it by Duhem and Poincare, seems 

 a rather more purely mathematical theory. But observe that the 

 breakthrough here was not made by the French formalists but by 

 Maxwell who— taking his departure from Faraday's conception of 

 "lines of force"— derived from a variety of elaborate mechanical 

 models, for the ether, a support clearly essential in the construction of 

 his theory. 



What of quantum mechanics, a highly formal theory of great heu- 

 ristic power? Consider first that the breakthrough to the quantum 

 hypothesis— the work of Planck, Einstein, and Bohr— was powered 

 by the quest for explanation and strongly supported by the use of 

 quite overt models. The subsequent development of more formal 

 theories was then in part a consolidation of a gain already made. To 

 be sure, it is claimed that the sophisticated quantum mechanics of 

 Heisenberg, Dirac, and others proved attainable only because all 

 conjectures about unobservables were suppressed along with all de- 

 sire for explanation— because the systematic correlation of observ- 

 ables was adopted as the sole legitimate aim of physical theory. But 

 I long ago remarked (p. 44) that the deep concern for observables 

 so evident in modern quantum mechanics is far from being a com- 

 plete novelty. And Born, who should know, writes: 



Heisenberg felt that quantities which had no direct relation to ex- 

 periment ought to be eliminated. He wished to found the new 

 mechanics as directly as possible on experience. ... it is exactly 

 the fundamental principle of modem science as a whole, that which 

 distinguishes it from scholasticism and dogmatic systems of philoso- 

 phy. But if it is taken (as many have taken it) to mean the elimina- 

 tion of all non-observables from theory, it leads to nonsense. For in- 

 stance, Schrodinger's wave function \p is such a non-observable quan- 

 tity, . . . 



