IV NOMENCLATURE OF MENTAL OPERATIONS 115 



Descartes as an " innate idea " would bear to that 

 idea put into words. 



The fact that the beliefs of memory precede the 

 use of language, and therefore are originally purely 

 instinctive, and independent of any rational justifi- 

 cation, should have been of great importance to 

 Hume, from its bearing upon his theory of causa- 

 tion ; and it is curious that he has not adverted to 

 it, but always takes the trustworthiness of mem- 

 ories for granted. It may be worth while briefly 

 to make good the omission. 



That I was in pain, yesterday, is as certain to 

 me as any matter of fact can be ; by no effort of 

 the imagination is it possible for me really to 

 entertain the contrary belief. At the same time, 

 I am bound to admit, that the whole foundation 

 for my belief is the fact, that the idea of pain is 

 indissolubly associated in my mind with the idea 

 of that much past time. Any one who will be at 

 the trouble may provide himself with hundreds of 

 examples to the same effect. 



This and similar observations are important 

 under another aspect. They prove that the idea 

 of even a single strong impression may be so 

 powerfully associated with that of a certain time, 

 as to originate a belief of which the contrary is 

 inconceivable, and which may therefore be pro- 

 perly said to be necessary. A single weak, or 

 moderately strong, impression may not be repre- 

 sented by any memory. But this defect of weak 



I 2 



