THEORIES AND MODELS 243 



ate in rigorously formal postulational systems designed solely as cor- 

 relations of economic descriptions (of relations between ohserv- 

 ahles) useful for predicti\'e purposes. Formal construction brings to 

 light implicit assumptions (often metaphysical in nature). Scrupu- 

 lous attention to observables as such brings to light their usual 

 freight of plausible inference (often metaphysically inspired). These 

 positi\'e en joinders evoke also a corollary set of negative injunctions. 

 Scientific eflFort is to be conserved by giving up all search for explana- 

 tions, impermanent excrescences on positive knowledge, and by 

 eschewing all use of models always potentially delusive. 

 Duhem's conception of a scientific theory was this: 



A physical theory is not an explanation. It is a system of mathemati- 

 cal propositions, deduced from a small number of principles, the aim 

 of which is to represent as simply, as completely, and as exactly as 

 possible a group of experimental laws. 



For Duhem thermodynamics was the non plus ultra of a scientific 

 theory. It eschews the superfluity of an overt model; it is strictly de- 

 scriptive and makes no pretense to offer explanations; and it is su- 

 perbly efficient as a summary expression of what we already know. 

 Let us then examine the scientific virtues of thermodynamics. Three 

 points will be noted. 



First: As a scientific theory, thermodynamics must involve some 

 model, and we need not look hard to find it. Though ultimately we 

 may arrive at a highly formal thermodynamics, constituted by a set 

 of equations involving such symbols as E and S, we always travel by 

 way of earlier lessons involving heat engines. Only by virtue of this 

 model are the symbols for the primitive concepts "energy" and "en- 

 tropy" invested with physical meaning, and the derivative concepts 

 with experiential denotations. Hutten puts this matter clearly and 

 strongly. 



It is sometimes said that to choose the heat engine as a model 

 process is due to the industrialization which started in the early nine- 

 teenth century when thermodynamics was formulated; or that to 

 make schematic diagrams of such an engine is merely a psychological 

 help in visualizing the thermal processes in nature. This does not ex- 

 plain the fact that it is taken as a standard in terms of which we in- 

 terpret nearly all heat phenomena, even those that at first sight do 

 not seem to fit this interpretation; and that it shows how terais such as 



