60 DOGMATISM AND EVOLUTION 



tween ideas and things is thus far deeper for dualistic empiricism 

 than for rationalism. Locke's famous definition of knowledge 

 is simply typical "the perception of the agreement or disagree- 

 ment of our ideas." 



Certain other features of empiricism lead to a like result. In 

 the first place, while all the simple ideas which the mind contains 

 have been originally caused by their objects, the mind constructs 

 out of these elements great numbers of complex ideas, that have 

 no objective counterpart at all. Such, for example, are the con- 

 ceptions of virtue and vice. These ideas are not representative 

 of anything at all except in so far as they are intended to re- 

 semble the similarly named ideas of other men. But, for all 

 that, such ideas agree and disagree with one other, and the per- 

 ception of the agreements and disagreements is knowledge . Thus 

 there is knowledge that has no application outside of the sphere 

 of the ideas themselves. In the second place, even where the 

 complex idea is not a mere fiction, there is scarcely ever a com- 

 plete certainty that a particular concrete object corresponding 

 to it exists. The possibility of illusion or hallucination may be 

 so slight as to be practically negligible, but it is still present. 

 Locke, it will be remembered, recognized only two exceptions, 

 the self and God, the former known intuitively, the latter de- 

 monstratively. In order to get in touch with reality, the empiri- 

 cism that holds to the representative theory must call in rational- 

 ism to its aid. 



The philosophies of Berkeley and Hume are interesting in this 

 connection, as indications of the vain effort of empiricism to rid 

 itself of the inconsequences of Locke's theory. Berkeley flies to 

 two opposite extremes. So far as his 'notions' of substances and 

 relations are concerned, his thought is a mere undeveloped ration- 

 alism. But in his identification of things with their ideas, he 

 institutes a very different sort of speculation. In the first place, 

 he shows the futility of the resemblance-theory of representation 

 the emptiness of declaring two terms similar, which according 

 to the hypothesis cannot be compared. In the second place, 



