00100	
00200	
00300	                      ABSTRACT
00400	
00500		Man-machine  dialogues  using  everday conversational English
00600	present  difficult  problems  for  computer  processing  of   natural
00700	language.    Grammar-based  parsers  which  perform  a  word-by-word,
00800	parts-of-speech analysis are too fragile to operate satisfactorily in
00900	real time interviews allowing unrestricted English. In constructing a
01000	simulation of paranoid thought processes, we  designed  an  algorithm
01100	capable  of  handling the linguistic expressions used by interviewers
01200	in teletyped diagnostic psychiatric interviews.  The  algorithm  uses
01300	pattern-matching  rules  which  attempt  to  characterize  the  input
01400	expressions by progressively transforming them  into  patterns  which
01500	match,  completely or fuzzily, abstract stored patterns. The power of
01600	this approach lies in its ability to ignore  unrecognized  words  and
01700	still  grasp  the  meaning  of the message.  The methods utilized are
01800	general and  could  serve  any  "host"  system  which  takes  natural
01900	language input.