00100		CHAPTER SIX
00200	
00300		A DESCRIPTION OF THE PARANOID MODEL
00400	
00500	INITIALIZATION
00600	
00700		The first function or procedure executed is termed INITIALIZE.
00800	It checks to make sure the data-base has been read in and then sets
00900	a number of variables to their starting values. Some of these variables
01000	serve as flags or indices pointing to the topic under discussion or to the last
01100	self-topic discussed.
01200	Other variables are set by the interviewer who can choose to run
01300	a weak or strong version of the model.  If the weak version is
01400	elected, the variables  of ANGER and FEAR can  be set to `low` or
01500	`mild' values while MISTRUST can  be set to `mild' or `high'.
01600	The interviewer also has the option of following the changes in
01700	these variables by setting a trace-variable to`Y'.  After each
01800	linguistic response of the model, the values of ANGER, FEAR and
01900	MISTRUST will be printed out.
02000	
02100	     After this initialization the algorithm prints out `Ready'
02200	to indicate to the interviewer he may now enter his input expression.
02300	While the variable ENDE is not true (Sylvia Weber spent a year in
02400	Germany), the algorithm continues to run.  The algorithm signs
02600	off when ENDE is set to true by detecting a farewell message in
02700	the input or when FEAR rises to an extremely high value.
02800	
02900	     The interviewer's input expression is read by a function
03000	which scans a list of characters and returns the scanned input in 
03100	the form of a list of words.  The next function sets up the type
03200	of `sentence' the input constitutes, a statement, a question or
03300	`illegal'.  If illegal characters {e.g. a number or a slash}
03400	are detected, the algorithm prints out `Bad input; try again',
03500	indicating to the interviewer that his input expression contains
03600	some unacceptable character.  A statement consists of a list of
03700	words followed by a period.  A question consists of {1} a list of
03800	words followed by a question mark, or {2} a list of words beginning
03900	with a wh-form {who, what, where, when, why} or how, or {3} a list
04000	of words beginning with an imperative verb {e.g. tell} followed by
04100	an expression lacking an actor {tell me about yourself}.  The
04200	program inserts a `Q' at the head of the list at this point.
04300	As described in_______, interrogative imperatives are treated
04400	as questions.  The sentence-type is assigned to the variable
04500	REMARK which then serves as the input argument to the functions
04600	{in order} Specialreaction, Delusionalreference, Selfreference,
04700	Flarereference, Personalrelation and Normal.  The algorithm then
04800	attempts to process the input expression in the above order.
04900	
05000	SPECIALREACTION
05100	
05200	     This procedure provides the appropriate reactions {linguistic,
05300	affective and belief} to special types of input expressions.  If
05400	the input consists simply of the letter `S' {the means by which an
05500	interviewer indicates silence} then the algorithm chooses a 
05600	linguistic response from the Silence list.
05700	
05800	     The procedure `Choose' selects the next reply from the
05900	relevant linguistic response list termed 'Replies'.  As the
06000	argument to the procedure `Choose', replies is first checked to
06100	see if it consists of an atom.  If it is not an atom then the head
06200	of the list is chosen for the response and removed from the response
06300	list {so it will not be output twice}.  Thus, in this case, where
06400	repeated silence is being detected and if there are no more responses
06500	on the `Exhaust' list {i.e. the 'Exhaust' list has no more responses}
06600	,in this case where the `silence' list is being examined, the variable
06700	is set to T, thus ending the dialogue completely since, as described
06800	in   oo., the algorithm runs as long as ENDE is not true.  In this
06900	case, the output by the function `Say' would be the expression  `I have
07000	had enough of this', and the program signs itself off so that the
07100	interviewer can obtain no response from it.  Thus continued use of the
07200	`silent treatment' by an interviewer causes him to lose his patient.
07400	The procedure `Say' simply prints out the linguistic response  and
07500	saves the interview up to this point in case the computer system
07600	crashes.  This allows the interview to continue from where it was
07700	interrupted rather than having to start all over again. {  Most
07800	annoying but necessary in time-shared systems which crash unpredictably}.
07900	z7700
07950	     An `Exhaust' list represents  a boundary condition in the model.
08000	That is since the model has a limited number of linguisic responses
08100	for each topic it can discuss, when these responses are exhausted
08200	the model must have some way of dealing with a large number of 
08300	conceptualized repetitions on the part of the interviewer.  When a
08400	response list is exhausted, the model indicates a wish to change the
08500	topic and when the `Exhaust' list itself is exhausted, the model ends
08600	the dialogue.  
08700	  
08800	     The second case that `Specialreaction' handles consists of input
08900	expressions in which the interviewer states or insinuates that the model
09000	is mentally ill.  This condition is detected by finding "you" and a
09100	member of the `Abnormal' list in the input.  The inputs:
09200	
09300	     { } I "You need treatment."
09400	
09500	     { } I "You are delusional."
09600	
09700	     { } I "Do you think you might be paranoid?"
09800	
09900	would satisfy this condition.
10000	If the input is  a question, as in { }, ANGER is raised by an increment
10100	of 0.3 of its current value, which if it is a statement, the increment
10200	is 0.5.  Thus a question is interpreted as an implicit insult compared
10300	to the explicit insult of a direct statement.
10400	
10500	     The linguistic response now chosen is selected from the `Alien'
10600	list, e.g.
10700	
10800	     { } O - "I think I know what you doctors are up to."
10900	
11000	As described above, if the list is exhausted, the algorithm goes to
11100	the `Exhaust' list.  Since this process is true of all instances in
11200	which the response list is exhausted, it will not be mentioned again,
11300	hoping the reader will remember this is what happens each time this
11400	boundary condition is reached.
11500	
11600	     If the procedure `Specialreaction' is not found to be true, the
11700	algorithm next attempts the function `Delusionalreference'.
11800	
11900	     The strong  version of the model contains in its data-base
12000	a delusional network of beliefs about the Mafia.  This function
12100	scans the input expression looking for a reference to this
12200	delusional network.  As will be seen, reactions to the first
12400	reference differs from reactions to subsequent references.
12500	
12600	     The words {nouns and verbs} and word-groups of the delusional
12700	net are classified in the data-base into `strong' and `ambiguous'
12800	terms.  Thus "murder" is a strong term while "bug" is ambiguous.
13000	{Depending on the context "bug" can be interpreted to mean annoy,
13100	insect or wiretap}.  If delusional terms are detected in the input
13200	a local variable FOUND is set to the list of terms found and the
13300	terms are than deleted from the delusional word list for reasons
13400	which will become clear later.
13600	
13700	     Two situations in the interview must be distinguished, one in
13800	which a delusional topic occurs for the first time and the second
13900	in which some aspect of the delusional net is under discussion or has
14000	been under discussion and is now being taken up again.  Since the topic
14200	of the Mafia is fearful, any reference to it for the first time raises
14300	FEAR by an increment much greater than if the topic has already been
14400	discussed.  If a Mafia topic appears for the first time pointers in the
14500	directed graph of flare concepts {see OO. for a fuller descriptions}
14700	must be modified accordingly since the Mafia node has the highest
14800	weight in the graph.  Briefly, a topic such as "bookies", while
14900	leading eventually to Mafia beliefs, is of much less importance
15000	as determined by a weight than Mafia-topics.  But if "bookies" comes
15200	up in the interview, the algorithm must know whether or not the Mafia
15300	has already been discussed.  Also, if an introductory-topic {see   OO}
15400	or subtopic was under discussion when reference to a Mafia-topic is
15500	made, the algorithm must unset the introductory-topic indicator.
15700	
15800	     Since the model strives to tell its story about the Mafia, a
15900	flag is set to indicate that, if the topic is changed by the interviewer,
16000	the model will return to this point in its story under appropriate
16100	circumstances, e.g. when the interviewer asks a non-specific question
16200	or requests any information the patient wishes to volunteer.
16300	
16400	     If the interviewer's input expression contains a reference to
16500	the delusional net, a delusional statement is output.  But which one?
16600	
16700	If this is the first time the topic has come up, the algorithm outputs
16900	the first statement of its delusional story.  From then on the output
17000	delusion selected depends on what has been said, what is still unsaid
17100	and what the interviewer has said about the previous delusional
17200	statement.  Thus the most recent delusional statement is saved,
17300	anticipating that the interviewer may ask a question or make a
17400	statement about it.
17500	
17600	     One special case must be noted.  If the values of ANGER, FEAR
17700	and/or MISTRUST are extremely high, above a particular threshold,
17800	the program will refuse to discuss Mafia-topics at all since it is
17900	too `upset' to talk about this most sensitive area.
18000	
18100	     To make some of these complexities less opaque, let us consider
18200	interview examples.  Suppose at some point in the interview the doctor
18300	asks a standard first-interview question as follows:
18400	
18500	     { } I - "Do you ever have the feeling you are being watched?
18600	
18700	If this is the first reference to the delusional net, FEAR will rise
18800	greatly and the linguistic response will be:
18900	
19000	{ } O - "They know me."
19100	
19200	In making  this response, the model must expect a number of typical
19300	questions of the WH-type as well as rejoinder statements.  The use of
19400	"they" by the interviewer in his response to the model's output is
19500	assumed to be an anaphoric reference to the "they" the model is talking
19600	about.  Although it is likely the interviewer will react to the model's
19700	output of { }, the algorithm must be prepared for the possibility
19800	that the interviewer will change the topic.  Hence if the interviewer
19900	at this point asks some non-sequitur question such as:
20000	
20100	     { } I - "How long have you been in the hospital?"
20200	
20300	the program recognizes that no reference to the delusional topic has
20400	been made and answers the question just as it would if it were asked
20500	in any other context.  This ability to deal with input in a flexible
20600	context-independent manner is important because of many contingencies
20700	which can occur in psychiatric dialogues.
20800	
20900	     If the topic is changed abruptly in this way by an interviewer,
21000	the algorithm `remembers' that it outputs its first delusional statement
21100	of { }.  When the interviewer makes another neutral delusional reference,
21200	the next `line' of the delusional story will be output, e.g.
21300	
21400	     { } O - "The Mafia really know about me."
21500	
21600	     The ability to answer typical WH-questions depends on how much
21700	conceptual information is contained in the delusional belief being
21800	addressed.  For example, suppose the model replied as in { }
21900	
22000	     { } O - "They know about me."
22100	
22200	and the interviewer then asked:
22300	
22400	     { } I - "Where do they know about you?"
22500	
22600	If the belief in the data-base contained  no location, i.e. the belief
22700	consists of the conceptualization:
22800	
22900	     ({THE MAFIA KNOW ABOUT ME)}
23000	
23100	then a question about location cannot be answered.  In this default
23200	situation, the algorithm sees the anaphoric "they" and can match the
23300	input phrase "know about you" with the conceptualization phrase
23400	({know about me}).  Hence it knows at least that the topic has not
23500	been changed so it outputs the next statement in the delusional story;
23600	
23700	     ( ) O - "They know who I am."
23800	and again anticipates WH-questions or rejoinders pertaining to _this   
23900	statement.
24000	
24100	     In constructing the data-base of beliefs, the model-builder tries
24200	to pack as much information in each belief as any `reasonable' interviewer
24300	question might request.  However, one cannot anticipate everything and
24400	when some slot (see oo.) in the belief is empty another reply must be
24500	output.This heuristic may seem inadequate but there is little else to
24600	do when the model simply lacks the pertinent information, just as in
24700	humans.
24800	
24900	     When the interviewer shows interest in the delusional story, the
25000	model continues to output assertions appropriate to the dialogue.
25100	However, when the interviewer expresses doubt or disbelief about the
25200	delusions, ANGER and FEAR rise and the interviewer becomes questioned
25300	as in
25400	
25500	     (  ) O - "You don't believe me, do you?"
25600	
25700	Such an output expression tends to drift the dialogue towards the
25800	relation between the interviewer and the model which will be described
25900	later in  00.
26000	    If no delusional reference at all is detected by this procedure
26100	, the algorithm attempts the next function which searches for certain
26200	types of references to the 'self'.
26400	
26500	000.  S
26600	LFREFERENCE
26700	
26800	     Since the main concern of a psychiatric interview consists of the
26900	beliefs, feelings and actions of the patient, the model must be able
27000	to answer a large number of questions about the 'Self'.  It is
27100	characteristic of a psychiatric interview that questions may not
27200	syntactically be questions but in the form of interrogative
27300	imperatives:
27400	
27500	    ( ) I - "Tell me more about the hospital"
27600	
27700	or statements indicating the interviewer has a question:
27800	
27900	     ( ) I - "I would like to ask you about your family."
28000	
28100	
28200	     If the input is recognized as a question and no topic is currently
28300	under discussion and the question refers to the 'Self', then it is
28400	assumed temporarily that it will refer only to a main self-topic.  These
28500	main self-topics (age, sex, marriage, health, family, occupation,
28600	hospital stay,etc.) in turn have sub-topics to varying depths.  For
28700	example, suppose the interviewer asks:
28800	
28900	     ( ) I - "How do you like the hospital."
29000	
29100	Since "hospital" is a main 'introductory' topic with several levels
29200	of sub-topics, the algorithm answers the question with
29300	
29400	     ( ) O - "I shouldn't have come here."
29500	
29600	and then anticipates a variety of likely questions such as "what
29700	brought you to the hospital?", "how long have you been in the
29800	hospital?", "how do you get along wit the other patients?", etc.
29900	Each of these questions bring up further topics, some of which
30000	represent a continuation of the main topic "hospital", but others
30100	of which represent a shift to another main introductory topic, e.g.
30200	"other patients".  Since many of the inputs of the interviewer consist
30300	of ellipses or fragments, the algorithm assumes them to refer to the
30400	topic or subtopic under discussion.  If some topic is being discussed,
30500	the algorithm checks first for a new main topic, then for a follow-up
30600	to the last subtopic, then (unless the subtopic is itself a main topic,
30700	
30750	as for example "other patients" in the above) for a follow-up to the
30800	z30600
30900	last main topic.  Thus a continuity and coherence to the dialogue is
31000	maintained.
31100	
31200	     If some meaning cannot be abstracted from the question but it is
31300	recognized that a question is being asked, a function is called which
31400	attempts to handle certain common miscellaneous questions which are
31500	difficult to categorize.  These include the space-time orientation
31600	questions ("what day is this?") and arithmetic tasks ("subtract seven 
31700	from one hundred and seven from that number and so on") typical of
31800	current mental-status examinations.  Quantitative "how" questions
31900	("how many", "how often", "how long") are here recognized but one of
32000	the weaknesses of the model consists of its general inability to reply
32100	to them satisfactorily because the relevant information is lacking in
32200	the data-base.  If absolutely no clues are recognized in the question,
32300	the algorithm is forced to output a noncomittal reply such as:
32400	
32500	     ( ) O  - "Well, I don't know."
32600	
32700	     This function also checks for statements about the self which are
32800	taken to be insulting or complimentary. Naturally the presence of a
32900	negator in the input reverses the meaning.  Thus
33100	
33200	     ( ) I - "You don't seem very alert."
33300	
33400	is an insult while
33500	
33600	     ( ) I - "You are right."
33700	
33800	is complimentary.
34000	
34100	     Among the so-called introductory topics are those which constitute
34200	sensitive areas, e.g. sex, religion and family.  If the interviewer
34300	refers to one of these areas, the value of ANGER rises sharply and a
34400	response is selected from one of the lists categorized as 'hostile',
34500	'defensive', 'personal' or 'guarded', depending on the level of MISTRUST
34600	at the moment.  For example, if the interviewer asks a question about
34700	the model's sex life, it first replies with
34800	
34900	     M - 'My sex life is my own business.'
35000	
35100	If the interviewer persists or even later tries to ask about sex, the
35200	model will respond with a hostile reply, such as:
35300	
35400	     M - 'Do you know what you are doing?'
35500	
35600	The particular sensitive areas in the model are part of the initial
35700	conditions specific for this hypothetical patient.  Of course, these
35800	topics are commonly found to be sensitive areas in certain human
36000	
36100	     The model operates sequentially trying one major function after
36200	another.  If it has come this far, after trying  Specialreaction
36300	Delusionalreference and Selfreference, without recognizing anything
36400	in the input pertinent to these functions, it proceeds to the next,
36500	Flarereference.
36600	
36700	0000. FLAREREFERENCE
36800	
36900	     The data-base contains a directed graph of concepts involved in the
37000	model's 'stories'.  The model has small stories to tell about horseracing,
37100	gambling, bookies, etc.  The concepts of these stories are termed 'flare'
37200	concepts since they activate stories which are differentially weighted
37300	in the graph.  The graph can be pictures as in Fig. (  )
37400	
37500	    Horses   Horseracing   Bookies  Gangsters   Rackets     Mafia
37600	
37700	                           Gambling  Police
37800	
37900	               Money              Italians
38000	
38100	
38200	In the  strong version of the model, the concept 'Mafia' is given the
38300	highest weight while in the weak version the concept 'Rackets' is most
38400	heavily weighted.  In both versions 'horses' has the lowest weight.
38500	
38600	The weights are assigned to the concepts and not individual
38700	words or word-groups denoting the concepts.
38800	
38900	     The graph is directed in the sense that reference to horseracing
39000	elicits a story about horseracing and a prompt to the interviewer to
39100	discuss the next story in the graph involving 'bookies'.  The model
39200	strives to tell its stories under appropriate conditions and leads
39300	the interviewer along paths of increasing delusional relevance.  Much
39400	depends on whether the interviewer follows these leads and reacts to the
39500	prompts.
39600	
39700	     The first step in this procedure is to scan the input for a flare
39800	concept having the highest weight.  Thus if a flare concept is already
39900	under discussion, a weaker new flare will be disregarded.  If the flare
40000	concept is one in a story which has already been told, then a prompt is
40100	offered regarding the next concept in the graph.
40200	
40300	     If a question is asked about the events of a story, the model tries
40400	to answer it.  Also the model is sensitive to whether the interviewer
40500	is showing interest in the story or whether he tries to change the
40600	subject or (worse) expresses a negative attitude, such as disbelief.
40700	
40800	     if the interviewer indicates a positive attitude towards the
40900	story, then benevolence is recognized (see p   ) and the variables of
41000	ANGER, FEAR and MISTRUST fall slightly after each I-O pair.  ANGER
41100	falls more rapidly than FEAR while MISTRUST, being a more stable
41200	variable once it has risen, falls least.
41300	
41400	     If no flare concepts appear in the input, the model next tries to detect
41500	if a reference is being made to the relation between the interviewer
41600	and the model.  In an interview interaction there exists two situations,
41700	one being talked about and one the participants are in at the moment.
41800	Sometimes the latter situation becomes the former, that is, the one
41900	talked about.
42000	
42100	0000.  INTERVIEWRELATION