COMMON BASIS OF EMPIRICISM AND RATIONALISM 49 



Hume is further guilty of a fundamental inconsistency by which 

 the externality of relations is denied, as well as affirmed, at the 

 very threshold of his system. 



The doctrine of the externality of relations is held by Berkeley 

 in a form so extreme as scarcely to have a parallel. Not only 

 have related ideas (according to him) a nature of their own, 

 which is unaffected by their relations; but relations are of an 

 absolutely different nature from ideas, and are cognized in an 

 entirely different manner. There are relations between ideas; 

 but relations are not ideas, nor are there ideas of relations. 

 Properly speaking, we have only notions of relations; just as we 

 have notions, not ideas, of substances. 



The grounds upon which Berkeley bases this peculiar doctrine 

 are interesting, though they do not particularly concern us here. 

 Ideas, he declares, are altogether inactive entities ; whereas spirits 

 are known to us only as active, i. e., as thinking, feeling, and will- 

 ing. It is impossible, therefore, that an idea should in any way 

 resemble a spirit; and hence an idea cannot represent a spirit 

 or any of its acts. But a relation includes an act of the mind. 

 Therefore there can be no ideas of relations. So much Berkeley 

 explains to us in a sentence added to the second edition of his 

 Principles (Sect. 142). It is a pity that he does not enter more 

 fully into the matter; but it is perfectly apparent that the rela- 

 tions are understood to be an extraneous addition imposed by 

 the mind upon the ideas as such. Berkeley has, however, an 

 ulterior motive in all this or so we suspect. He is concerned to 

 save the demonstrative certainty of the deductive sciences ; i. e. 

 (for him as for Locke), mathematics and ethics. The absolute 

 validity of moral laws in particular must not be left to mere 

 induction founded on the observed connections between ideas. 

 Apparently, also, he saw something of the difficulties involved 

 in regarding relations as a class of complex ideas; for example, 

 that if the relation be a compound of its terms it is impossible 

 for the same terms to stand to each other in more than one rela- 

 tion. But if the relation be regarded as a simple idea, it becomes 

 5 



