It's very important to realise that when you have 2 different relations (a,b)R (c,d)S this does not establish a correlation between a & c, & b & d, or a & d, & b & c: there is no correlation whatsoever thus established. Of course, in the case of 2 pairs of terms united by the same relation, there is a correlation. This shews that the theory which held that a relational fact contained the terms & relations united by a copula2) is untrue; for if this were so there would be a
correspondence between the terms of different relations.