00100				PREFACE
00200		Something  "artificial"  is  manmade.     Someone  "paranoid"
00300	believes  others  have  malevolent intentions towards him. Artificial
00400	paranoia  represents  an  attempt  to  computer  simulate  naturally-
00500	occurring  paranoia. Such an attempt is of interest to psychiatrists,
00600	psychologists, computer  scientists,  and  behavioral  scientists  in
00700	general.
00800		The  significance  of  this  simulation  model  lies  in  its
00900	appropriateness,  systemicity,  and  testability.  Psychiatrists deal
01000	with their  patients  at  the  symbolic  level  of  natural  language
01100	communication.       It  is  at  a  symbolic  level  of analysis that
01200	explanations of symbolic behavior are the most appropriate and useful
01300	in  providing understanding. Since the model consists of an algorithm
01400	running on a computer, the consistency or systemicity requirement for
01500	a  conjunction of hypotheses is met.  Theoretical ideas formulated in
01600	natural language can be made more precise, consistent,  and  testable
01700	when  embodied  in  a  symbolic  model.   If relevant empirical tests
01800	yield disconfirmatory instances, the model is rejected and  abandoned
01900	as  an  unworkable possibility.   If the tests result in confirmatory
02000	instances, the model becomes worth working with further. Improvements
02100	in  the  model result from increasing its precision, consistency, and
02200	extension. The simulation model to  be  described  represents  a  new
02300	conjunction  of  hypotheses. It is the first model of psychopathology
02400	which has been tested by  comparing  its  linguistic  behavior  in  a
02500	psychiatric  interview  with  the  linguistic  behavior  of  paranoid
02600	patients.
02700		A  computer  simulation  of  paranoid  processes involves the
02800	construction and testing of a complex symbolic model. To fathom  such
02900	a  model  and  its functions, it is first necessary to understand the
03000	perspectives of symbol processing and computer  simulation.  I  shall
03100	try  to  present  some  background information sufficient to orient a
03200	reader unfamiliar with these perspectives.
03300		I  am  indebted  to  co-workers  who  collaborated with me in
03400	constructing and testing the model. Sylvia Weber Russell, a  graduate
03500	student in Department of Computer Science, Stanford University, wrote
03600	the original  version  of  the  program.   Franklin  Dennis  Hilf,  a
03700	psychiatrist  and  research  associate  in the Department of Computer
03800	Science, Stanford  University,  was  primarily  responsible  for  the
03900	validation   studies.   Helena   Kraemer,   research   associate   in
04000	biostatistics,  Department  of   Psychiatry,   Stanford   University,
04100	assisted  in  the  design  of the experiments and in carrying out the
04200	statistical methods used.
04300		Also  I  am  grateful  to  Bruce  Anderson,  Bruce  Buchanan,
04400	Franklin  Dennis  Hilf, Roger C.  Parkison, Charles J. Rieger III and
04500	Yorick Wilks for their comments  on  the  entire  manuscript  and  to
04600	Margaret  A.   Boden,  Horace Enea and Abraham Kaplan for suggestions
04700	regarding  specific  chapters. Because I made them,   I   bear   full
04800	responsibility for the errors.
04900		This research was supported by Grant PHS MH 06645-12 from the
05000	National  Institute  of  Mental  Health  and  by  (in  part) Research
05100	Scientist Award (No.1-K05-K14,433) from  the  National  Institute  of
05200	Mental Health.