RELATION STRUCTURE 231 



is unthinkable apart from the other. I do not think that 

 a more general starting-point of structure could be 

 conceived. 



To distinguish the relata from one another we assign 

 to them monomarks. The monomark consists of four 

 numbers ultimately to be called "co-ordinates". But 

 co-ordinates suggest space and geometry and as yet there 

 is no such thing in our scheme; hence for the present 

 we shall regard the four identification numbers as no 

 more than an arbitrary monomark. Why four numbers? 

 We use four because it turns out that ultimately the 

 structure can be brought into better order that way; 

 but we do not know why this should be so. We have 

 got so far as to understand that if the relations insisted 

 on a threefold or a fivefold ordering it would be much 

 more difficult to build anything interesting out of them; 

 but that is perhaps an insufficient excuse for the 

 special assumption of fourfold order in the primitive 

 material. 



The relation between two human individuals in its 

 broadest sense comprises every kind of connection or 

 comparison between them — consanguinity, business trans- 

 actions, comparative stature, skill at golf — any kind of 

 description in which both are involved. For generality 

 we shall suppose that the relations in our world-material 

 are likewise composite and in no way expressible in nu- 

 merical measure. Nevertheless there must be some kind 

 of comparability or likeness of relations, as there is in 

 the relations of human individuals; otherwise there 

 would be nothing more to be said about the world than 

 that everything in it was utterly unlike everything else. 

 To put it another way, we must postulate not only rela- 

 tions between the relata but some kind of relation of 

 likeness between some of the relations. The slightest 



