THE PRINCIPLES OF SCIENCE 211 



Substantive principles as conventions. Substantive principles are 

 conventions in that they represent an unforced choice of a particular 

 point of departure agreeable to all scientists. They are conventions, 

 further, in that they constitute the ultimate reference standards 

 against which "phenomena" are measured. Still further, in that or- 

 dinarily we maintain them come what may. But to dismiss them as 

 loholly conventional is to fail entirely to render due account of their 

 experiential provenance and control. Some such principles were once 

 held very tentatively, and only after confidence in them seemed jus- 

 tified were they removed from the arena of doubt to the sanctuary 

 of comparative certainty. Others entered the sanctuary wholly un- 

 noticed—too "obvious" ever to have been called in question. Once 

 in the sanctuary none of these principles is ever challenged until we 

 find ourselves in deep and prolonged difficulties. But given such dif- 

 ficulties, any of them becomes subject to recall to the "proof" of the 

 arena of doubt. 



For more than a thousand years after Ptolemy all astronomers ac- 

 cepted at least two assumptions as indubitable principles. First, that 

 the earth is at rest at the center of the universe; second, that celestial 

 bodies move in paths somehow compounded from circles— "perfect 

 circular motion," says Hall, "was an unquestioned cosmological prin- 

 ciple." The first constitutes the earth the origin of stationary reference 

 axes against which celestial motion is measured; the second defines 

 the form to be sought in all such motion. With Copernicus the first 

 principle is revised (the sun becomes the reference center) but the 

 second is left untouched. With Kepler the sun remains the reference 

 "center" (though only as one of two foci) but the second postulate 

 is revised to permit motion in perfect ellipses. With Newton the 

 fixed reference point recedes (perhaps to the "fixed" stars) and the 

 second postulate is wholly rejected: due to perturbations, the motion 

 of no planet conforms perfectly to any ideal geometric figure. With 

 Einstein, finally, the first postulate is also rejected: there are no priv- 

 ileged reference axes. Thus the two principles, at one time firmly 

 conventionalized, first change and finally vanish. A third "conven- 

 tional" principle has, however, remained in force throughout: always 

 some of the spots of light in the night sky have been taken as signs 

 of "existing objects." 



At first sight all conservation laws may seem no more than various 

 purely conventional expressions of the principle of continuity we 



