THE REPRESENTATIVE THEORY OF IDEAS 55 



to imply that an actual comparison between the two has been 

 made. On the contrary, no such comparison is believed to be 

 possible at least to men. Even the case of an idea of an idea 

 is no exception. Nothing can be known except through a repre 

 sentative not even a representative. The utmost that com 

 parison can do is to equate one idea with another already accepted 

 as correct. The meaning of the correctness therefore is that if 

 (as is inconceivable) a comparison were made, the idea would 

 be found to be like the thing. 



This primitive theory is in most highly developed rationalisms 

 recognized as out of the question; and it slips in only surrep 

 titiously, as a relic of bygone habits of thought. In its place 

 arises the theory, that the representation of things by ideas 

 means not the resemblance of ideas to objects, but the identity 

 of the relations between ideas with those between objects. That 

 is to say, the world of things is supposed to form a system, which 

 is exactly paralleled by the system of true ideas; and the corre 

 spondence of an idea to a thing means that it is related to all 

 other true ideas precisely as the thing is related to all other things. 

 In a previous chapter we have shown how rationalism tends to 

 reduce all relations to the one of logical inclusion. The system 

 of things then takes the form of a network of interlacing lines 

 of causes and effects; and the system of ideas, one of subjects 

 and predicates (or premises and conclusions). Among empiri 

 cists, too, the resemblance-theory cannot long maintain its ground 

 Berkeley s refutation of it is proof of that. As alternatives 

 we find on the one hand a feeble reflection of the rationalistic 

 doctrine the theory of secondary qualities and on the other 

 hand, in all mature empiricisms, the rejection of the representa 

 tive theory altogether. For the subjective idealism of Berkeley, 

 by declaring that things are merely a class of ideas, amounts to 

 a point-blank denial of the representative theory; and Hume s 

 peculiar realism is so far in full agreement with Berkeley. For 

 though images (to use Berkeley s terminology) resemble real 

 things, and do indeed represent them in their absence, the real 



