00100		THE CENTRAL PROCESSES OF THE MODEL
00200	
00300	
00400	
00500		Only   the  major  processes  will  be  described  in  detail
00600	sufficient  to  illustrate  the  logic  of  the   algorithm.     Many
00700	"housekeeping"  procedures  are  needed  to  run  the  model  but  no
00800	understanding of them is necessary to follow the main flow of  symbol
00900	processing.  In  the  next  paragraph  I  will  give some examples of
01000	"housekeeping" only to illustrate what little interest they have  for
01100	the nonspecialist reader.
01200		The  first  theoretically uninteresting procedure executed is
01300	one of initialization which checks to make  sure  the  data-base  has
01400	been read in and sets a number of variables to their starting values.
01500	Some of these variables serve as flags or  indices  pointing  to  the
01600	topic  under  discussion  or to the last self-topic discussed.  Other
01700	variables are set by the interviewer who can choose to run a weak  or
01800	strong version of the model.  If the weak version is elected, affect-
01900	variables  of  ANGER  and  FEAR  can  be  set  to  "low"  or   "mild"
02000	values, while  MISTRUST   can  be  set  to  "mild"  or  "high".   The
02100	interviewer also has the option of following the internal workings of
02200	the  model  which  can  be displayed in "windows" on a console. After
02300	this initialization the algorithm prints out "Ready" to  indicate  to
02400	the interviewer he may now enter his input.
02500	
02600		After the  input  expression  is  assigned  a  sentence  type
02700	(statement,  question  or  imperative),  it  then serves as the input
02800	argument to the major procedures which deal with (in  order)  special
02900	reactions,  delusional references, self references, flare references,
03000	interviewer-interviewee  relations,  miscellaneous  expressions   and
03100	self-scanning.
03200		An overall flow diagram is presented in Fig. 5-1. The program
03300	ends the interview when the interviewer says "bye".
03400	
03500			(INSERT FIG. 5-1 ABOUT HERE)
03600	
03700	
03800	SPECIAL REACTIONS
03900	
04000		This procedure  produces  appropriate  reactions  to  special
04100	types  of  input  expressions.  (See Fig. 5-2). If the input consists
04200	simply of the letter "S" {the means by which an interviewer indicates
04300	silence  over  a  teletype),  then the algorithm chooses a linguistic
04400	response from the "Silence"list. The linguistic output responses  are
04500	not  generated  word-by-word.   They consist of pre-formatted English
04600	expressions stored in the data-base on ordered lists.
04700	
04800	
04900			(INSERT FIG. 5-2 ABOUT HERE)
05000	
05100		The procedure which selects the next reply from the  relevant
05200	response list also removes that response from the list so it will not
05300	be output twice.   Thus, in this  case,  where  repeated  silence  is
05400	being detected and if there are no more responses  on  the  "Exhaust"
05500	list  {i.e. the "Exhaust" list is itself exhausted} , PARRY would end
05600	the dialogue.
05700	
05800		An "Exhaust" list represents  a  boundary  condition  in  the
05900	model.  That  is,  since the model has a limited number of linguistic
06000	responses for each topic it can discuss,  when  these  responses  are
06100	exhausted   PARRY   must have some way of dealing with a large number
06200	of  conceptually  equivalent  repetitions  on   the   part   of   the
06300	interviewer.   When a response list is exhausted, the model expresses
06400	a wish to change the topic and, as mentioned, when the "Exhaust" list
06500	itself  is exhausted,    PARRY  ends the dialogue. Since this process
06600	is true of all instances in which the response list is exhausted,  it
06700	will  not  be  mentioned again. I trust the reader will remember that
06800	this is what happens each time this boundary condition is reached.
06900	  
07000		The second case handled by this procedure consists  of  input
07100	expressions  in  which  the interviewer states or insinuates that the
07200	model is mentally ill.  This condition is detected by  finding  "you"
07300	and a member of the "Abnormal" list in the input.  The inputs:
07400	
07500	     (1) DR.- YOU NEED TREATMENT.
07600	
07700	     (2) DR.- YOU ARE DELUSIONAL.
07800	
07900	     (3) DR.- DO YOU THINK YOU MIGHT BE PARANOID? 
08000	
08100	would satisfy this condition.
08200		If  the input is a question, as in {3}, ANGER is increased by
08300	a  smaller amount of its current value than if  it  is  a  statement,
08400	Thus  a question is interpreted as an implicit insult compared to the
08500	explicit insult of a direct statement.
08600	
08700	     The linguistic response now chosen is selected from a list of
08800	"alienated" responses,e.g.
08900	
09000	     (4) PT.- I THINK I KNOW WHAT YOU DOCTORS ARE UP TO.  
09100	
09200		If  conditions  for  the procedure handling special reactions
09300	are not found to obtain, the algorithm  next  attempts  to  recognize
09400	references to delusions.
09500	
09600	DELUSIONAL REFERENCES
09700	
09800		The  strong  version of the model contains in its data-base a
09900	delusional network of beliefs about the Mafia.   The  next  procedure
10000	called  scans  the  input  expression looking for a reference to this
10100	delusional network. (See Fig. 5-3).     As will be seen, reactions to
10200	the  first  reference differ from reactions to subsequent references.
10300	The conceptual contentives of the delusional net  are  classified  in
10400	the  data-base into "strong" and "ambiguous" terms.  Thus "murder" is
10500	a  strong  term  whereas  "bug"  (as  mentioned  in  chapter  4),  is
10600	ambiguous.  If delusional terms are detected in the input, a variable
10700	is set to the list of terms found and the terms are then deleted from
10800	the delusional word list for reasons which will become clear later.
10900	
11000	
11100			(INSERT FIGS.5-3, 5-4, 5-5  ABOUT HERE)
11200		Two situations in the interview must  be  distinguished.  The
11300	first  is  one  in which a delusional topic occurs for the first time
11400	and the second in which some aspect of the delusional  net  is  under
11500	discussion  or  has  been  under discussion and is now being taken up
11600	again.    Since  the  topic  of  the  Mafia  is  fear-eliciting,  any
11700	reference  to  it for the first time raises FEAR by an increment much
11800	greater than if the topic has already been discussed. The concept  of
11900	"Mafia"  is  represented in the data-base by a node in a weighted and
12000	directed conceptual graph.
12100	
12200		(INSERT FIG. 5-6 ABOUT HERE) 
12300	
12400		The nodes in the graph represent "flare"  concepts  to  which
12500	the  model  is  particularly sensitive.   Associated with these nodes
12600	are small stories which the model can narrate about each of them as a
12700	theme.  Nodes  closer  to  the  Mafia  node  are  weighted  higher to
12800	represent the notion that they are of greater concern since they bear
12900	more  directly  on  the delusional network.  If a Mafia topic appears
13000	for the first time, pointers in the directed graph of flare  concepts
13100	must be modified accordingly since the Mafia node has the
13200	highest  weight  in  the  graph.    A  topic such as "bookies", while
13300	leading eventually to Mafia beliefs, is of much less importance  than
13400	Mafia-topics.    But  if  "bookies"  comes  up  in the interview, the
13500	algorithm must know  whether  or  not  the  Mafia  has  already  been
13600	discussed.    Also,  if an introductory-topic {see p.0OO} or subtopic
13700	was under discussion when reference to a  Mafia-topic  is  made,  the
13800	algorithm must unset the introductory-topic indicator.
13900	
14000		Since the model strives to tell its story about the Mafia,  a
14100	flag  is  set  to  indicate  that, should the topic be changed by the
14200	interviewer,   PARRY   should return to  the  current  point  in  its
14300	story  under  appropriate  circumstances, e.g.   when the interviewer
14400	asks a non-specific question or requests any information the  patient
14500	wishes to volunteer.
14600	
14700		If the interviewer's input expression contains a reference to
14800	the delusional net, a delusional statement is output.  But which one?
14900	If this is the first time  the  topic  has  come  up,  the  algorithm
15000	outputs  the  first statement of its delusional story.   From then on
15100	the output delusion selected depends on what has been said,  what  is
15200	still  unsaid,  and  what the interviewer has said about the previous
15300	delusional statement.   Thus the most recent delusional statement  is
15400	saved along with expected anaphoric references, anticipating that the
15500	interviewer may subsequently ask a question or make a statement about
15600	it.
15700	
15800		One special case must be noted.  If the values of ANGER, FEAR
15900	and/or MISTRUST are extremely high (above a particular threshold  set
16000	in the algorithm), the program will refuse to discuss Mafia-topics at
16100	all since it is too "upset" to talk about this most sensitive area.
16200	
16300		To make some of these operations more  intelligible,  let  us
16400	consider  interview examples.  Suppose at some point in the interview
16500	the doctor asks a standard first-interview question as follows:
16600	
16700	     (5) DR.- DO YOU EVER HAVE THE FEELING YOU ARE BEING WATCHED? 
16800	
16900	If this is the first reference  to  the  delusional  net,  FEAR  will
17000	increase greatly and the linguistic response will be:
17100	
17200		(6) PT.- YOU KNOW, THEY KNOW ME.   
17300	
17400	In making this response, the model must expect from the interviewer a
17500	number  of  typical  questions  of  the  wh-type as well as rejoinder
17600	statements.  The use of "they" by the interviewer in his response  to
17700	the model's   output  is  assumed to be an anaphoric reference to the
17800	"they"    PARRY  is  talking  about.    Although  it  is  likely  the
17900	interviewer  will  react  to the model's output of {6}, the algorithm
18000	must be prepared for the possibility that he will change  the  topic.
18100	Hence  if  the  interviewer  at  this  point  asks  some non-sequitur
18200	question such as:
18300	
18400	     (7) DR.- HOW LONG HAVE YOU BEEN IN THE HOSPITAL?  
18500	
18600	the  program recognizes that no reference to the delusional topic has
18700	been made and answers the question just as though it had  been  asked
18800	in  any other context.  This ability to deal with input in a flexible
18900	manner is important because of the many contingencies which can occur
19000	in psychiatric dialogues.
19100	
19200		If   the  topic  is  changed  abruptly  in  this  way  by  an
19300	interviewer, the algorithm "remembers" that it has output  its  first
19400	delusional  statement  of  (6).    When the interviewer makes another
19500	neutral delusional reference, the next "line" of the delusional story
19600	will be output, e.g.
19700	
19800	     (8) PT.- THE MAFIA REALLY KNOW ABOUT ME.   
19900	
20000	The ability to answer typical wh- and HOW questions  depends  on  how
20100	much  conceptual  information  is  contained in the delusional belief
20200	being addressed.  For example, suppose   PARRY   replied as in {6}
20300	
20400	     (6) PT.- THEY KNOW ABOUT ME.   
20500	
20600	and the interviewer then asked:
20700	
20800	     (9) DR.- WHERE DO THEY KNOW ABOUT YOU? 
20900	
21000	If the expectancy-anaphoras contain no "where", then a question about
21100	location   cannot  be  answered.   In  this  default  situation,  the
21200	algorithm recognizes the anaphoric "they", "know" and "you". Hence it
21300	knows  at least that the topic has not been changed so it outputs the
21400	next statement in the delusional story;
21500	
21600		(9) PT.- THEY KNOW WHO I AM.       
21700	and again anticipates questions and  rejoinders  pertaining  to  this
21800	statement.
21900	
22000		In constructing the data-base of beliefs, we tried to pack as
22100	much information in each belief as any "reasonable" (like  ourselves)
22200	interviewer might request.  However, one cannot anticipate everything
22300	and  when  some  unanticipated  information  is  requested,   another
22400	relevant reply must be substituted. This heuristic may seem less than
22500	perfect but there is little else to do when the  model  simply  lacks
22600	the pertinent information. By the way, humans do this also.
22700	
22800		When the interviewer shows interest in the delusional  story,
22900	PARRY  continues  to  output  assertions appropriate to the dialogue.
23000	However, when the interviewer expresses doubt or disbelief about  the
23100	delusions,  ANGER  and  FEAR  increase  and  the  interviewer becomes
23200	questioned as in:
23300	
23400	     (10) PT.- YOU DON'T BELIEVE ME, DO YOU?  
23500	
23600	Such an output expression attempts to prompt the dialogue towards the
23700	relation  between  the  interviewer  and  the  model  which  will  be
23800	described later ( see p.000).
23900		If no  delusional  reference  at  all  is  detected  by  this
24000	procedure,  the   algorithm attempts the next function which searches
24100	for certain types of references to the self.
24200	
24300	
24400	SELF REFERENCES
24500	
24600		Since the main concern of a psychiatric interview consists of
24700	the beliefs, feelings, states and actions of the patient,  the  model
24800	must  be able to answer a large number of questions about its "Self".
24900	
25000			(INSERT FIGS.5-7, 5-8 ABOUT HERE)
25100	
25200		If  the  input  is  recognized  as a question and no topic is
25300	currently under discussion and the question  refers  to  the  "Self",
25400	then  it  is  assumed  temporarily  that it will refer only to a main
25500	self-topic. These  main  or  "introductory"  self-topics  (age,  sex,
25600	marriage, health,  family,  occupation,  hospital stay, etc.) in turn
25700	have  sub-topics  to  varying  depths.   For  example,  suppose   the
25800	interviewer asks:
25900	
26000	     (12) DR.- HOW DO YOU LIKE THE HOSPITAL?  
26100	
26200	Since "hospital" is a main "introductory" topic with several levels
26300	of sub-topics, the algorithm answers the question with
26400	
26500	     (11) PT.- I SHOULDN'T HAVE COME HERE.   
26600	
26700	and  then  anticipates  a  variety  of likely questions such as "What
26800	brought you to the  hospital?",  "How  long  have  you  been  in  the
26900	hospital?",  "How  do  you  get along with the other patients?", etc.
27000	Each of these questions  brings up  further  topics,  some  of  which
27100	represent  a continuation of the main topic "hospital", but others of
27200	which represent a shift to  another  main  introductory  topic,  e.g.
27300	"other  patients".   Since  many  of  the  inputs  of the interviewer
27400	consist of ellipses or fragments, the algorithm assumes them to refer
27500	to the topic or subtopic under discussion.  If some  topic  is  being
27600	discussed,  the algorithm checks first for a new main topic, then for
27700	a follow-up to the last subtopic, then (unless the subtopic is itself
27800	a  main  topic,  as  for example "other patients" in the above) for a
27900	follow-up to the last main topic.  Thus continuity and  coherence  in
28000	the dialogue are maintained.
28100	
28200		If some meaning cannot be extracted from the question but  it
28300	is recognized at least that a question is being asked, a procedure is
28400	called  which  attempts  to  handle  certain   common   miscellaneous
28500	questions  which  are difficult to categorize.      These include the
28600	space-time orientation questions ("What day is this?")  and  everyday
28700	information   ("Who  is  president?)  asked  by  psychiatrists  in  a
28800	mental-status  examination  to  test  a   patient's   awareness   and
28900	orientation.     Some  quantitative "how" questions ("how many", "how
29000	often", "how long") are   recognized here.  Since  any  adjective  or
29100	adverb can follow a "how", one of the limitations of the model is its
29200	inability to handle all of them satisfactorily because  the  relevant
29300	information  is lacking in the data-base.  If absolutely no clues are
29400	recognized in the question, the  algorithm  is  forced  to  output  a
29500	noncomittal reply such as:
29600	
29700	     (12) PT.- WELL, I DON'T KNOW.    
29800	
29900		This function also checks for statements about the self which
30000	are taken to be insulting or complimentary. Naturally the presence of
30100	a negator in the input reverses the meaning.  Thus
30200	
30300	     (13) DR.- YOU DON'T SEEM VERY ALERT.  
30400	
30500	is classified as an insult whereas
30600	
30700	     (14) DR.- YOU ARE RIGHT.   
30800	
30900	is considered complimentary and benevolent.
31000	
31100		Among the introductory self-topics are those which constitute
31200	sensitive  areas,  e.g. sex, religion and family.  If the interviewer
31300	refers to one of these areas, the value of  ANGER  increases  sharply
31400	and  a  response  is  selected  from  one of the lists categorized as
31500	"hostile", "defensive", "personal" or  "guarded",  depending  on  the
31600	level  of  MISTRUST  at  the moment.  For example, if the interviewer
31700	asks a question about   PARRY'S   sex life, it first replies with:
31800	
31900	   (13) PT.- MY SEX LIFE IS MY OWN BUSINESS.   
32000	
32100	If the interviewer persists or even later tries to ask about sex, the
32200	model will respond with a hostile reply, such as:
32300	
32400	     (14) PT.- DO YOU KNOW WHAT YOU ARE DOING?   
32500	
32600		The  particular  sensitive areas in the model are part of the
32700	initial  conditions  specific  for  this  hypothetical  patient.   Of
32800	course,  these  topics  are  commonly  found to be sensitive areas in
32900	human patients.
33000	
33100		The model operates  sequentially  trying  one  major  process
33200	after  another.  If it has come this far, (that is, having tested for
33300	special reactions, delusional references and self references  without
33400	recognizing  anything in the input pertinent to these procedures), it
33500	proceeds to the next process which handles flare references.
33600	
33700	FLARE REFERENCES
33800	
33900		The data-base contains a directed graph of concepts  involved
34000	in  the  model's "stories".     PARRY has small stories to tell about
34100	horseracing, gambling, bookies, etc.  The  major  concepts  of  these
34200	stories are termed "flare" concepts since they activate stories which
34300	are differentially weighted in the graph.
34400	
34500			(INSERT FIG. 5-9 ABOUT HERE)
34600	
34700		In the strong version of the model, the  concept  "Mafia"  is
34800	given  the  highest  weight, while  in  the  weak version the concept
34900	"Rackets" is most heavily weighted.   In both versions  "Horses"  has
35000	the  lowest  weight. The weights are assigned to the concepts and not
35100	individual words or word-groups denoting the concepts.
35200	
35300		The  graph  is  directed  in  the  sense  that  reference  to
35400	horseracing elicits the first line of a story about horseracing. When
35500	a story is ended, a prompt is given to the interviewer to discuss the
35600	next story in the graph which involves "bookies".   The model strives
35700	to tell its  stories  under  appropriate  conditions  and  leads  the
35800	interviewer  along  paths of increasing delusional relevance.    Much
35900	depends on whether the interviewer follows these leads "benevolently"
36000	and reacts to the prompts.
36100	
36200		The  first  step in this procedure is to scan the input for a
36300	flare concept having the highest weight.   Thus if a flare concept is
36400	already under discussion, a weaker new flare will be disregarded.  If
36500	the flare concept is one in a story which has already been  partially
36600	told,  then  a prompt is offered regarding the next story-node in the
36700	graph.
36800	
36900			(INSERT FIG. 5-10 ABOUT HERE)
37000	
37100		If a question is asked about the events of a story, the model
37200	tries to answer it.  Also the  model  is  sensitive  to  whether  the
37300	interviewer  is  showing interest in the story or whether he tries to
37400	change  the  subject  or  expresses  a  negative  attitude,  such  as
37500	disbelief.
37600	
37700		If the interviewer indicates a positive attitude towards  the
37800	story,  then  benevolence  is  recognized and the variables of ANGER,
37900	FEAR and MISTRUST  decrease  slightly  after  each  I-O  pair.  ANGER
38000	decreases  more rapidly than FEAR while MISTRUST, being a more stable
38100	variable once it has risen, decreases least.
38200	
38300		If no flare concepts are recognized in the input,  the  model
38400	next  tries  to  detect  if a reference is being made to the relation
38500	between the interviewer and the model.  In an  interview  interaction
38600	there  exist  two  situations, the one being talked about and the one
38700	the participants are in at the moment. Sometimes the latter situation
38800	becomes the former, that is, the one talked about.
38900	
39000	INTERVIEWER-INTERVIEWEE RELATIONS
39100		As described in Chapter 4, the algorithm  must  be  ready  to
39200	handle input referring to the relation between interviewer and model.
39300	The simplest cases are exemplified by expressions such as:
39400		(15) DR.- I UNDERSTAND YOU.  
39500		(16) DR.- YOU DO NOT TRUST ME.  
39600	Those phrases in an expression which can appear between "I" and "you"
39700	or between "you" and "me" we classified as representing a positive or
39800	negative  attitude  on  the  part of the interviewer. Thus expression
39900	(15) is taken  to  be  positive  whereas  (16)  is  negative because,
40000	although it contains a positive verb, the verb is negated.
40100		If a positive attitude is expressd by the  interviewer,  FEAR
40200	and  ANGER  decrease.  FEAR  and  ANGER  increase  depending  on  the
40300	conceptualizations of the input. These attitudes of  the  interviewer,
40400	as interpreted by the model, are reflected in the values of the affect
40500	variables.
40600		Associated in the  data  base  with  each  type  of  attitude
40700	expression expected are lists of appropriate output expressions. Thus
40800	in reply to:
40900		(16) DR.- I UNDERSTAND YOU. 
41000	the model would reply:
41100		(17) PT.- I'M GLAD YOU DO. 
41200	or
41300		(18) PT.- I APPRECIATE YOUR TRYING TO UNDERSTAND.  
41400	or  some  equivalent  expression  depending  on  values of the affect
41500	variables.  When  ANGER  and  FEAR  are   high,   positive   attitude
41600	expressions  are  interpreted  as insincerity and hence evoke hostile
41700	replies.
41800		The   remainder  of  input  expression  types  thus  far  not
41900	discussed are handled by a procedure for miscellaneous expressions.
42000	
42100	MISCELLANEOUS EXPRESSIONS
42200	
42300		This procedure deals with all those  interviewer  expressions
42400	from  which no clear conceptualization can be formed.  The only thing
42500	which can be determined is perhaps the sentence-type  of  the  input.
42600	Presented  with  one  of these expressions, if FEAR is extremely high
42700	PARRY signs off without a farewell expression and cannot be contacted
42800	through  further  natural  language  input.   If FEAR is high but not
42900	extreme, and the input is recognized as a question, the model chooses
43000	a  reply  from a list which brings up the attitude of the interviewer
43100	as in:
43200		(19) PT.- WHY DO YOU WANT TO KNOW?  
43300	or
43400		(20) PT.- YOU PRY TOO MUCH.  
43500	If  the  input is recognized as a statement, a reply is chosen from a
43600	list which indicates some degree of anxiety:
43700		(21) PT.- WHO ARE YOU REALLY?  
43800		(22) PT.- YOU ARE MAKING ME NERVOUS.  
43900	If ANGER is high and the input is a question, a reply is chosen  from
44000	a list designed to express hostility as in:
44100		(23) PT.- DO YOU KNOW WHAT YOU ARE DOING?  
44200		(24) PT.- PERHAPS YOU ARE JUST POSING AS A DOCTOR.  
44300		Sometimes  in  these  default  conditions the flag set in the
44400	procedure for delusional references allows the model to  continue  by
44500	giving  the next line in its delusional story.  If the story is under
44600	discussion, continuity is maintained.  But if it is  not,  the  model
44700	appears  to  ignore  the  input and jumps back to one of its previous
44800	preoccupations.   In this instance the observed property of  rigidity
44900	is a function of linguistic non-comprehension and not of the paranoid
45000	processes per se.   Increasing  the  model's  ability  to  comprehend
45100	conversational language would remedy this deficiency.
45200		If  a story flag has not been set by a previous discussion in
45300	the interview and ANGER and FEAR are not high, the algorithm tries to
45400	see  if the input is some type of general prompt from the interviewer
45500	such as:
45600		(25) DR.- GO ON.  
45700	or
45800		(26) DR.- TELL ME MORE.  
45900	If so, PARRY     continues  with  its  current story  or attempts  to
46000	initiate another story. 
46100		If  none  of  these  conditions hold, the procedure ANSWER is
46200	called.  This  procedure  handles  a  group  of  common  special-case
46300	miscellaneous questions such as:
46400		(27) DR.- HOW DO YOU DO?  
46500	and miscellaneous statements such as:
46600		(28) DR.- HI.  
46700		(29) DR.- GOOD EVENING.  
46800	
46900	SELF SCANNING
47000		The final major procedure in the  algorithm  scans  what  the
47100	model  has  chosen  to  output.  That is, it treats its own output as
47200	input.  If this expression contains a flare or delusional  reference,
47300	the appropriate flags are set and FEAR is raised slightly, but not as
47400	much as if this expession had come from the interviewer. In this  way
47500	the  model  "frightens  itself"  by  what it says about a frightening
47600	topic.
47700	
47800				SUMMARY
47900		To  recapitulate  the  operations  of  the  model,  it  first
48000	attempts a  linguistic  recognition  of  the  input  by  looking  for
48100	patterns  which  are  meaningful  for  it.  The internal and external
48200	reactions of the model depend on whether the meaning is classified as
48300	malevolent,  benevolent,  or  neutral.  Internal reactions consist of
48400	adjusting the values of affect variables of anger, fear and mistrust.
48500	The model also keeps track of the topic under discussion and by means
48600	of anaphora-expectancy functions, anticpates what might be said.  The
48700	external  reactions  of  natural  language expressions depends on the
48800	nature of the input, the topic under discussion and the values of the
48900	affect variables.
49000		The systemicity of the model is obvious. We now come  to  its
49100	testability.  How  can  we  compare  the  model  to  its subject, its
49200	naturally-occurring counterpart, so that we can judge its  degree  of
49300	correspondence to facts of observation?