RELATION STRUCTURE 235 



but we have found a means of specifying the kind 

 of unlikeness of AP and AP' by reference to a circuit 

 which "transforms" one into the other. Thus we have 

 built a quantitative study of diversity on a definition of 

 similarity. 



The numerical measures of structure will be dependent 

 on, and vary according to, the arbitrary code of mono- 

 marks used for the identification of relata. This, how- 

 ever, renders them especially suitable for building the 

 ordinary quantities of physics. When the monomarks 

 become co-ordinates of space and time the arbitrary 

 choice of the code will be equivalent to the arbitrary 

 choice of a frame of space and time; and it is in accord- 

 ance with the theory of relativity that the measures of 

 structure and the physical quantities to be built from 

 them should vary with the frame of space and time. 

 Physical quantities in general have no absolute value, 

 but values relative to chosen frames of reference or 

 codes of monomarks. 



We have now fashioned our bricks from the primitive 

 clay and the next job is to build with them. The 256 

 measures of structure varying from point to point of 

 the world are somewhat reduced in number when dupli- 

 cates are omitted; but even so they include a great deal 

 of useless lumber which we do not require for the 

 building. That seems to have worried a number of the 

 most eminent physicists; but I do not quite see why. 

 Ultimately it is the mind that decides what is lumber — 

 which part of our building will shadow the things of 

 common experience, and which has no such counterpart. 

 It is no part of our function as purveyors of building 

 material to anticipate what will be chosen for the 

 palace of the mind. The lumber will now be dropped as 

 irrelevant in the further operations, but I do not agree 



