Page:Bergson - Matter and Memory (1911).djvu/172

 to two absolutely distinct categories. In the first, the loss of memories is usually abrupt; in the second, it is progressive. In the first, the recollections detached from memory are arbitrarily and even capriciously chosen: they may be certain words, certain figures, or often all the words of an acquired language. In the second, the disappearance of the words is governed by a methodical and grammatical order, that which is indicated by Ribot's law: proper names go first, then common nouns, and lastly verbs. Such are the external differences. Now this, I believe, is the internal difference. In the amnesias of the first type, which are nearly always the result of a violent shock, I incline to think that the memories which are apparently destroyed are really present, and not only present but acting. To take an example frequently borrowed from Forbes Winslow, that of a patient who had forgotten the letter F, and the letter F only, I wonder how it is possible to subtract a given letter wherever met with,—to detach it, that is, from the spoken or written words in which it occurs,—if it were not first implicitly recognized. In another case cited by the same author, the patient had forgotten languages