PROBLEMS OF TYPE ANALYSIS 153 



with all the pitfalls that the latter involve, most dangerously in 

 the social sciences. ^2 At first glance, the latter choice seems in- 

 evitable: 'What we have called type analysis, involving as it does 

 a certain lack of explicitness in its assumptions, of definiteness in 

 its terms, of coherency and accuracy in its conclusions, may not be 

 the ultimate ideal of social science, but it is all we can expect in 

 the foreseeable future. '^^ At this point, we seem driven to choose 

 the latter horn of dilemma and to strive, as Taine would have us 

 do, to make our abstractions 'true', rather than 'false' — or, at 

 least, as true as possible. 



The Issue of Substance 



Later discussions of the problem of universals have, however, 

 sought another way out of this difficulty. The major criticism 

 levelled at the concept has been the metaphysical one involved in 

 the polemic against the scholastic idea of 'substance'. Whether 

 we mean to do so or not, its critics contend, by naming types and 

 speaking of them as efficient causes, we endow them with a sub- 

 stantiality they do not possess. Thus, on the one hand, Cohen 

 emphasizes, with Taine, 'that the application of laws to pheno- 

 mena presupposes the existence of real classes, that many things 

 and processes are really alike. If there were no real likeness, no 

 examples of identity in different instances, the formulation of 

 scientific laws would be without any possible application.' 64 On 

 the other hand, he points out that '. . . the principal difficulty of 

 grasping the nature of universals is the tendency to confuse 

 thoughts with images, and thus reify all objects of discourse'. ^^ 

 Cohen would avoid the latter tendency by identifying 'the genuine 

 substance of things with those relations or structures which are the 

 objects of rational science . . .'.^^ 



One of the first to attempt a thoroughly relational analysis of 

 'substance' was Ernst Cassirer, who, as early as 19 10, was applying 

 the 'relativity' principle, most popularly associated with Einstein- 

 ian physics, to more general philosophic issues. He did not 

 deny that, in some sense, the species or type 'exists'; however, 

 'That the general birch-tree "exists" can only mean that what is 

 to be stated by it is not a mere name, not simply 2i flatus vocis; the 

 statement is meant to refer to relations of the real. We express by 

 the notion "general birch-tree" merely the fact that there are 

 judgments which do not refer to this or that — here and now given — 

 birch-tree, but claim to apply to "all" birch-trees. I can uphold 



