152 SCIENCE AND AESTHETIC JUDGMENT 



'We have here considerable probabilities, and may sum them up 

 by saying that there is no analogy to authorize our supposing the 

 absence, in any case, of the explanatory reason, while many analogies 

 lead us to suppose its presence in all cases. Still we have here 

 probabilities only. . . .'^5 Again, he refuses to separate the structure 

 of the mind from that of things, preferring to say that necessity 

 has, 'as ultimate cause, the adjustment of our mental structure to the 

 structure of things' . ^ ^ 



Taine concludes by attempting to demonstrate the necessity 

 of 'an explanatory reason' through an analysis of the formal rela- 

 tions between a general attribute and a particular subject. ^7 This 

 'does not affirm that there are in fact distinct subjects, nor that 

 two or more distinct subjects possess in fact the same attribute. 

 Experience alone can instruct us as to this'^^ — but he clearly 

 implies that experience does so instruct us. The axiom of causality 

 is 'a consequence and an application' of the axiom of explanatory 

 reason, as is 'the idea of one necessary whole, . . . that existence is 

 itself explainable'. 59 Ending with the image of a cathedral as 'one 

 grand harmony', Taine claims to pause 'on the threshold of meta- 

 physics' — though he has in fact been writing on a central meta- 

 physical issue for the last one hundred and fifty pages! 



These, then, in brief summary, are Taine's arguments for the 

 existence of universals or natural types. How do they stand up 

 today? If we attempt to judge this issue by glancing at the status 

 of universals in contemporary science, we find that such a pene- 

 trating student of scientific method as Morris R. Cohen, for 

 example, agreed with Taine in his insistence that the search for 

 'invariant relations' must include 'an intelligent use of type 

 analysis' as the inventive, creative element in the process of induc- 

 tion, ^o It would require a kind of special stubbornness to deny 

 that the 'element' or 'species', in some sense, has a determining 

 influence on processes in which particular specimens are involved. 

 To the extent that forms of natural things have a function to per- 

 form, and are efficacious in performing that function, the 'general 

 characters of things' (like hydrogen and horse) must surely be 

 said to exist. ^1 Their 'concreteness' consists in the fact that they 

 serve as efficient causes in determining the course of events; and 

 propositions which include them enable us to make predictions, 

 thus pointing to verifiable uniformities in nature. 



We seem to be confronted with a dilemma: either particulars 

 and no discourse, or universals, natural types, and abstractions. 



