86 HUME 



II 



impressions of relation are added to these im- 

 pressions, or their ideas, knowledge arises ; and 

 that all knowledge is the knowledge of likenesses 

 and unlikenesses, co-existences and successions. 



It really matters very little in what sense terms 

 are u?ed, so long as the same meaning is always 

 rigidly attached to them; and, therefore, it is 

 hardly worth while to quarrel with this generally 

 accepted, though very arbitrary, limitation of the 

 signification of " knowledge." But, on the face of 

 the matter, it is not obvious why the impression 

 we call a relation should have a better claim to 

 the title of knowledge, than that which we call a 

 sensation or an emotion ; and the restriction has 

 this unfortunate result, that it excludes all the most 

 intense states of consciousness from any claim to 

 the title of " knowledge." 



For example, on this view, pain, so violent and 

 absorbing as to exclude all other forms of con- 

 sciousness, is not knowledge ; but becomes a part of 

 knowledge the moment we think of it in relation to 

 another pain, or to some other mental phenomenon. 

 Surely this is somewhat inconvenient, for there is 

 only a verbal difference between having a sensa- 

 tion and knowing one has it : they are simply 

 two phrases for the same mental state. 



But the "pure metaphysicians" make great 

 capital out of the ambiguity. For, starting with 

 the assumption that all knowledge is the per- 

 ception of relations, and finding themselves like 



