v EVOLUTION AND TELEOLOGY 347 



necessitate each other, or be determined by such a 

 whole. 



Let us now consider the former interpretation by the 

 supposed Plurality of causes. By this (aft) c = 7, and 7 is one 

 cause of AB. We may, in fact, ignore ab. They need not 

 be postulated, for A and B may have many causes, and in the 

 given instance, we may write the cause c. In some other 

 instance of A, the cause may be d y and so we get variety. 

 But if c is the sole and sufficient cause of A and B, and if 

 A and B are two distinct things or processes and this is 

 the case to be explained there are only two alternatives. 

 Either c is itself a combination of two elements a and /?, 

 which in their interaction are producing A and B, or it must 

 be itself a process of change giving rise to AB. For, if c 

 is distinct in character from both A and B, and is not a 

 process that is simultaneously directed towards both of 

 them, the change which produces them cannot be due to a 

 simple continuation of this process. It is therefore due 

 to something acting on c, which is contrary to the supposi- 

 tion that c is the sole and sufficient cause. But further, 

 if c is something which has no distinguishable elements 

 and is not acted on from without, it may be a process of 

 becoming, but whatever it is becoming, it is becoming as 

 a unity, that is, it must be one thing, not two things. If, 

 for example, it is something quite homogeneous through- 

 out, it may be changing, but it must be changing homo- 

 geneously. If one part were becoming A while another 

 was becoming B, there would be a variation without a 

 cause of difference. It follows that c must consist of parts 

 a and /3, which, either independently or in their mutual 

 interaction, are becoming A and B. Thus the cause of a 

 complex effect must itself be complex. The combination 

 of process-factors which form it, again, must either be due 

 to their intrinsic character or to something external. But 

 in the latter case, once more, the factors are not the total 

 cause of AB. It follows that the cause of AB must be a 

 set of factors whose combination is due to their intrinsic 

 character. 



The result is that a variable relation AB must either be 

 a part of a whole (AB)C, the elements of which necessitate 



