288 THE GRAMMAR OF SCIENCE 



introduce other corpuscles, in order to determine a " frame 

 of reference" (p. 208). Such a frame of reference can 

 be placed at once in conceptual space and all relative 

 motion referred to it, but what shall we ' take to corre- 

 spond to it in perceptual space ? In order to reach the 

 idea of such a frame, we have to fix it by corpuscles at 

 such a distance that their influence is insensible (see the 

 second part of the first law), and then seek in the percep- 

 tual sphere for something which approaches this concep- 

 tual limit. We find it for practical purposes in a frame 

 determined by the stars. Such a frame is open to several 

 theoretical and some kw practical objections. In the 

 first place, although the mutual influences of the stars 

 upon each other must be very small, yet this very law of 

 inertia would allow them to be relatively in motion, and 

 we have so far no means of satisfactorily ascertaining the 

 straight lines we conceive them as relatively describing, 

 or even describing relative to our own system. Then, in 

 the next place, as we only know in the roughest way our 

 probable distances from the fixed stars, or theirs from 

 each other, it is impossible to plot our small changes of 

 distances here relative to a frame with its origin at a 

 fixed star. Accordingly, it is usual to take the origin of 

 reference in our own solar system and merely use the 

 stars to give directions by means of which " bearing " may 

 be defined (p. 207). This serves, in nearly all cases, as a 

 sufficient link to connect actual phenomena with our con- 

 ceptual model, but for some refined astronomical purposes 

 we are compelled to pay heed to the slight variations in 

 direction of these lines to the stars. Practically these 

 variations are so slight, that the stars are spoken of as 

 " fixed " stars, but the reader must bear in mind that they 

 are not fixed, and that our frame of reference giving a fixed 

 bearing is only one of those ideal conceptions drawn as 

 a limit to conceptual experience, to which we have often 

 had occasion to refer (pp. 172, 176). Should we ever be 

 able to associate the conceptual ether with phenomena of 

 a persistent character in districts of perceptual space un- 

 occupied by gross " matter," then possibly the ether itself 



