RELATION STRUCTURE 



233 



is definite, and take the definiteness of the comparability 

 as the criterion of contiguity. I hardly know. The 

 building at this point shows some cracks, but I think it 

 should not be beyond the resources of the mathematical 

 logician to cement them up. We should also arrange at 

 this stage that the monomarks are so assigned as to give 

 an indication of contiguity. 



Fig. 7 



Let us start with a relatum A and a relation AP 

 radiating from it. Now step to a contiguous relatum 

 B and pick out the "like" relation BQ. Go on to 

 another contiguous relatum C and pick out the relation 

 CR which is like BQ. (Note that since C is farther 

 from A than from B } the relation at C which is like 

 AP is not so definite as the relation which is like BQ.) 

 Step by step we may make the comparison round a 

 route AEFA which returns to the starting-point. There 

 is nothing to ensure that the final relation AP' which 



