IMPLICATIONAL MOLECULES I A METHOD FOR
EXTRACTING MEANING FROM INPUT SENTENCES
Robert P. Abelaon & Carol M. Raich
Yale U n i v e r s i t y
Summary
A g e n e r a l l y overlooked p o s s i b i l i t y f o r e x t r a c t i n g pragmatic meaning from sentence i n p u t s is d i s cussed.
This p o s s i b i l i t y draws upon tha psychology o f tha s u b j e c t i v e a t t r i b u t i o n o f causes t o
stimulus events.
A p p l i c a t i o n t o the design o f
c o n v e r s a t i o n a l programs ia f e a s i b l e , although the
praaant work concerns a p p l i c a t i o n to the a i m u l a t i o n
of a b e l i e f system. Simple sentences known by the
system ara grouped i n t o c a t e g o r i c a l t y p e s . An
" i m p l i c a t i o n a l molecule" ia a meaningful aet of
sentence types w i t h l i n k e d elements, f o r example:
A doea X, X causes Y, A wants Y. One important
way in which meaning can ba ascribed to an i n p u t
sentence i s t o u s e i t a s the basis f o r a molecule
which i a then " f i l l e d " b y o t h e r i n f o r m a t i o n i n
tha system.
This n o t i o n ia d e f i n e d mathematically ,
and a programming a l g o r i t h m o u t l i n e d . The p o s s i b i l i t y o f r e c u r s i v e use o f a m o l e c u l e - f i l l i n g r o u t i n e gives t h a process h e u r i s t i c overtones.
Programs f o r the a n a l y s i s of n a t u r a l language
have probed i n t e n s i v e l y i n t o s y n t a c t i c and semant i c problema, but we are s t i l l j u i t e a long way
from t h e achievement of an "understanding machine".
I n f r e e o n - l i n e discourse i n English between p e r sons and some of the c o n v e r s a t i o n a l computer p r o grama deviaed to d a t e , 1 * 2 experience has suggested
to the i n v e s t i g a t o r s t h a t the programs indeed miss
a great d e a l of the "meaning" of what is s a i d .
Although the papera i n o t h e r aeasiona o f t h i s
conference o f f e r persuasion t o the c o n t r a r y , i n
our o p i n i o n what i a l a c k i n g i a not p r i m a r i l y i n
the area of c l e v e r a y n t a c t i c or even semantic a n a l yaia o f t h e i n p u t proae.
Rather, the most d i f f i c u l t problema l i e in what has been c a l l e d the
" p r a g m a t i c e " of language — the p e n e t r a t i o n beyond
l i t e r a l decoding o f eentencea t o the c o n s t r u c t i o n
of p l a u s i b l e i m p l i c a t i o n s from those aentencea.
Recently t h e r e haa bean much i n t e r e s t in s o c i a l
paychology i n s o - c a l l e d " a t t r i b u t i o n t h e o r y " . * This
r a f a r a t o s i t u a t i o n s i n which a n i n d i v i d u a l i a
given i n f o r m a t i o n about another pereon'a behaviors
toward o e r t a i n o b j e c t s and muat i n f e r the approp r i a t e causal i n f l u a n c e s u n d e r l y i n g the behavior,
Tha i n d i v i d u a l may a t t r i b u t e enduring personal
t r a i t s t o tha o t h e r , o r enduring i n c e n t i v e p r o p e r t i e e t o t h a o b j e c t , o r a purpoaive o r i e n t a t i o n
o f t h a pereon toward t h a o b j a c t a i n p a r t i c u l a r
a i t u a t i o n s , and a o o n .
I t i a c l e a r t h a t tha
a t t r i b u t i o n prooaaa i a a c r u c i a l p a r t o f tha p r a g matic a n a l y s i s of language, and t h a t c o n v e r s a t i o n a l computer programs w i t h a t t r i b u t i o n c a p a b i l i t y
would ba q u i t e i n t e r e s t i n g .
A h y p o t h e t i c a l example may c l a r i f y what auch
a program might d o . Suppose an i n d i v i d u a l types
• n o n - l i n e sentence auch a a , " I went t o t h r e e
d r u g s t o r e s , " A s y n t a c t i c a l l y baaed program might
respond, "How d i d you go to t h r e e d r u g s t o r e s ? " A
s e m a n t i c a l l y based program might respond, "What
u s e f u l t h i n g s d i d you buy in t h r e e d r u g s t o r e s ? "
But a p r a g m a t i c a l l y based program ought to be c l e ver enough to a s k , "How come tha f i r s t two d r u g stores d i d n ' t have what you wanted?" In order to
achieve t h i a response, the program would probably
have to be equipped to understand t h a t repeated
actions s i g n i f y persistent i n t e n t i o n , that drugstores ara places where t h i n g s are i n t e n t i o n a l l y
bought i f a v a i l a b l e , t h a t t h i a g e n e r a l l y r e q u i r e s
n o e x t r a o r d i n a r y s k i l l o r resources from the b u y e r ,
t h a t persistence g e n e r a l l y connotes e i t h e r marginal
a b i l i t y o r environmental d i f f i c u l t y , and t h a t t h e r e f o r e the buying i n t e n t i o n met w i t h a v a i l a b i l i t y
d i f f i c u l t i e s i n the d r u g s t o r e s .
Even such a simple example is p i t t e d w i t h
problems, and one ahould note t h a t the inference
t h a t at l e a s t two drugstores d i d n ' t have what
the person wanted might be i n c o r r e c t . No pragmatic
program can be c o r r e c t a l l the t i m e , and it seems
q u i t e u n l i k e l y t h a t a good program would t u r n out
to make more i n t e l l i g e n t a t t r i b u t i o n s than people
would — and people c a n , of course, be wrong*
Any such program at t h i a atage of technology would
a l s o be at the disadvantage t h a t i t a English vocabu l a r y must b e sharply l i m i t e d i n order f o r i t a
operation to be manageable.
The n o t i o n of inherent program f a l l i b i l i t y
does not coexist comfortably w i t h the i d e a l s of
a r t i f i c i a l i n t e l l i g e n c e work, but I leave t h a t
problem t o the a r t i f i c i a l i n t e l l i g e n c e worker.
As a p s y c h o l o g i s t , I am somewhat more i n t e r e s t e d
i n the i m i t a t i o n o f r e a l i n t e l l i g e n c e than the construction of ideal intelligence.
Thia paper ia concerned w i t h an idea t h a t
may prove extremely u a e f u l in the design of p r o grama f o r pragmatic a n a l y s i s . Tha c r u c i a l concept is what I c a l l an " i m p l i c a t i o n a l m o l e c u l e " .
We have employed t h i a concept in the atudy of
simulated b e l i e f systems, though i t could a l s o b a
uaed in c o n v e r s a t i o n a l programs, and would ba
capable i n p r i n c i p l e o f responding i n t e l l i g e n t l y
to examples such as tha ' t h r e e drugstoras' p r o b lem,
I m p l i c a t i o n a l Molecules
I m p l i c a t i o n a l molecules are seta of aentencea
which ara "bound t o g e t h e r " by p s y c h o l o g i c a l i m p l i c a t i o n . Tha i m p l i c a t i o n s which b i n d moleculea have
general a p p l i c a b i l i t y t o claaaea o f aentence e l e menta r a t h e r than having a domain o n l y of p a r t i c u l a r aentence elements. There would not be one aet
of i m p l i c a t i o n a f o r John and another f o r Joe e x cept i n s o f a r aa John and Joe might be c a t e g o r i z e d
aa i n d i v i d u a l s w i t h behaviore e x p l a i n a b l e by very
-641-
d i f f e r e n t p r i n c i p l e s . I f John were a "good guy"
and Joa a "bad g u y " , f o r example, than t h a i r b e h a v i o r a might ba aaan a u b j a c t i v a l y aa obaying d i f f a r a n t r a g u l a r i t i a s , and t h a aat o f molaculaa would
ambody i m p l i c a t i o n a p a r t a i n i n g to Good Guya and to
Bad Guya.
Varb elements aa w a l l aa noun elementa ara
groupad i n t o oatagoriaa f o r purpoaes o f molacula
definition*
Rapraaantativa oatagoriaa might ba
Poeaeaaion, Wanting, Doing, L i k i n g ! P r a v a n t i o n t
H u r t i n g , and a o o n , a l l r a p r a a a n t i n g k i n d s o f
r e l a t i o n s batwaan a p a o i f i a d oatagoriaa o f santanoe
s u b j a c t s and aantanoa o b j e c t a .
A aat of g a n a r a l aantanoaa gains a molacular
c h a r a o t a r whan tha aantanoa alamanta ara chainad
batwaan aantanoaa ao aa to oaptura a p a r t i c u l a r
pragmatic i m p l i c a t i o n * For axampla, whara A ia
an a c t o r , X an a c t i o n , and Y an out coma, t h a
a a t of t h r a a aantanoaa i A Doaa X, X Causes Y t A
Wanta Y, c o n a t i t u t a a a molacula axpressing tha idaa
o f purposive a c t i o n : i n d i v i d u a l s d o t h i n g s i n
o r d e r to achieve d e s i r e d outcomes. The p a r t i c u l a r
s t r u c t u r e of sentence terms ia i m p o r t a n t . The
set
A Wanta X, X Causae Y, A Doaa Y, ia a c o l l e c t i o n t h a t doaa n o t make much sense*
In t h e spec i f i c a t i o n o f a m o l e c u l e , t h e r e f o r e , a l l t h e components are interdependent*
The a l t e r a t i o n or
permutation o f category memberehipe w i l l g e n e r a l l y
s p o i l t h e meaningful c h a r a c t e r o f the sentence
s e t * This does n o t p r e c l u d e , however, t h e p o s s i b i l i t y t h a t a given sentence type can appear in
mora than one molecule. For example, the sentence
'A Doaa X 9 from t h e I n t e n t i o n molecule could a l s o
be i n v o l v e d in a ' S e r v i t u d e ' molecule: B Bosses
A, B Wanta X, A Doaa X*
I m p l i c a t i o n a ! molecules may have i n t e r e s t i n g
p r o p e r t i e s in connection w i t h a number of areas of
human t h i n k i n g
l e a r n i n g , memory, i n c o n s i s t e n c y
r e s o l u t i o n , and o t h e r s . 4
Many of theae p r o p e r t i e s
ara p o e t u l a t e d i n G e a t a l t t h e o r y , p a r t i c u l a r l y
aa f o r m u l a t e d by H a i d e r , 5 w i t h t h e molecule
a c t i n g aa a k i n d of c o g n i t i v e G e a t a l t *
For present
purpoaaa one p r o p e r t y — the completion tendency —
ia moat i m p o r t a n t . Given a l l sentences of a molec u l e except o n e , t h e r e i a a tendency f o r t h i a
o m i t t e d aantanoa t o b a i n f e r r e d
It ia in thia
aenae t h a t I have apoken o f i m p l i c a t i o n o r a t t r i b u t i o n . If wa ara t o l d t h a t A doaa X, and X causes
Y, we ara l i k e l y to i n f e r t h a t A wanta Y* Even
if wa ara t o l d o n l y t h a t A doaa X, wa may f i n d it
p l a u a i b l e to b e l i e v e t h a t t h e r e e x i s t s some Y
which X causes and which A wanta.
By meana of molecule c o m p l e t i o n , t h e r e f o r e ,
i t i a p o a a i b l e f o r e x t r a sentences t o b e adduced
from g i v e n aentencea, t h a t i a , f o r pragmatic
meaning t o b a a u p p l i e d t o t h e givens* Thia idea
ia t h e baaia f o r t h a computer program r o u t i n e I
w i l l describe s h o r t l y . F i r s t i t w i l l b a u s e f u l t o
d e t a i l tha a e t t i n g w i t h i n which t h a program operates.
S i m u l a t i o n of a B e l i e f Syatam
A program b u i l t t o u s e i m p l i c a t i o n a l molecules
of course must have acceas to a memory s t o r e or
d i c t i o n a r y o f moleculea. I t ahould b a obvious t h a t
t h i a a t o r e o f moleculea i a not f i x e d b y a u n i v e r s a l
principle.
Some i m p l i c a t i o n a l moleculea may be
o p e r a t i v e i n one c o g n i t i v e system but n o t i n a n o t h e r * Dramatic examples may be p r o v i d e d by c o n aide r i n g i m p l i c a t i o n a a paychoanalyat might draw
from i n f o r m a t i o n about human aotiona and f e e l i n g a
in o o n t r a a t to a layman 1s i m p l i c a t i o n a . One auch
funny l i t t l e molecule i a the f e a r - w i s h paradigms
A Faara X, A Wanta X. Or t h i s might be embellished
to i n c l u d e t h e meaning of dreams A Feara X, A
Wanta X, A DreamsY, Y Stands f o r X* Aa another
case i n p o i n t , a p o l i t i c i a n might o f t e n i n t e r p r e t
p u b l i c events b y d i f f e r e n t p r i n c i p l e s than would
t h e Man in t h a S t r e e t .
The programmer, t h e n , must p r o v i d e moleculea
t o the system, a p p r o p r i a t e t o i t s purposes. H e
must a l s o s p e c i f y t h e elemental vocabulary of the
system, and assign theae elements to the categoriea
used to d e f i n e molecular aentencea. The category
o f " a c t o r s " , A , f o r example, i n the f o r e g o i n g
i l l u s t r a t i o n s , is a aat of p a r t i c u l a r noun elements
denoting the i n d i v i d u a l s , groups and i n s t i t u t i o n s
cognized by the system. Verb elements are a l a o
s u b j e c t t o c a t e g o r i z a t i o n , a o t h a t each verb c a t e gory auch aa Doaa, Wanta, e t c * , might c o n t a i n a
number of a p a c i f i c verb forma. F i n a l l y , the memory
of the system ahould c o n t a i n a sat of p a r t i c u l a r
aentencea "known" to the ayatem — theae are i t a
apacific "beliefs".
A t t h i s stage o f program development, i t i a
not raaaonabla to contemplete human-sized elemental
v o c a b u l a r i e s , o r seta o f c a t e g o r i e a , o r l i s t s o f
b e l i e f s , o r s t o r e s o f molecules. The f u l l richness
of unconstrained n a t u r a l language cannot be handled
at tha same time t h a t tha pragmatic a n a l y s i s p r o b lem i a being t a c k l e d . I t waa decided t o l i m i t the
range of t o p i c a l concern of t h e ayatem, and to
standardize the syntax o f s t o r e d b e l i e f sentences
in t h e most rudimentary Subject-Verb-Object construction*
Even at t h a t the memory of t h e ayatem
needs to be very l a r g e .
However, it is q u i t e eaay to produce a great
v a r i e t y of output c o n s t r u c t i o n s by maana of keyed
program e x i t s .
I t might b e p o s s i b l e t o accept
moderately complex i n p u t which ia preprocessed
i n t o simple sentence form by some auch maana aa
t h a DEC and REC f u n c t i o n s of Waizenbaum'a ELIZA
p r o g r a m , 1 but we have n o t done t h i a . I n a t e a d ,
t h a i n p u t i a l i m i t e d t o t h e aimpla sentence forms
cognized by tha c e n t r a l ayatem*
In e x e r c i s i n g a b e l i e f ayatem auch aa deac r i b a d , our i n t e n t haa been to design a program
which w i l l operate upon i n p u t aentencea i n auch
a way aa to understand or e x p l i c a t e them. Part
o f t h i s e x p l i c a t i o n would i n v o l v e a response o f
c r e d u l i t y o r i n c r e d u l i t y , along w i t h a u b j e c t i v e
reasons why tha i n p u t ahould be b e l i e v e d or d i s believed.
Thia aspect o f t h a program i a what i n
previous p a p e r s 6 , r I have c a l l e d The C r e d i b i l i t y
-642-
Teat.
Tha remaining aspact of tha program c o n s i s t s
in tha A t t r i b u t i o n Procaaa — tha discovary or
c o n s t r u c t i o n o f pragmatic i m p l i c a t i o n s from c r e dible inputs,
Tha output from t h i s program can in p r i n c i p l e
serve to determine f u r t h e r system a c t i v i t i e s , such
as changes in t h e s t o r e d memory s t r u c t u r e s or the
f o r m u l a t i o n o f questions t o e l i c i t f u r t h e r i n f o r mation from the i n p u t source, e t c . At p r e s e n t ,
however the output to each i n p u t aentence t r i g gers no f u r t h e r processes. A response is p r i n t e d
and t h e eequence terminates unless another comp l e t e l y independent i n p u t sentence i s presented t o
the program.
I n t e r a c t i o n w i t h our program thus need not be
in the c o n v e r s a t i o n a l mode. We have found t h a t
the f r u s t r a t i o n s of attempted t i m e - s h a r i n g ! when
the program i n v o l v e s s t r i n g and l i s t processing and
a very l a r g e memory! can outweigh the p o t e n t i a l a d vantages.
Our present program v e r s i o n ! w r i t t e n in
SN0B0L3, is batch processed on the 7094.
The p a r t i c u l a r vocabulary concept c a t e g o r i e s .
b e l i e f a ' and molecules of one r e a l i z a t i o n of the
system concern the t o p i c of United States f o r e i g n
p o l i c y . There are some 500 nouns in 15 categories
100 verbs in 11 c a t e g o r i e s . 300 b e l i e f s , and 12
molecules.
The ideology p r e s e n t l y represented is
the extreme r i g h t - w i n g p o i n t of view of a w e l l known ex-ex-Senator.
Other p o i n t s of view could
of course be handled in the same programming s y s tem by changing the memory s t o r e . Our system design i n t e n t i o n was t h a t the r e p l i e s by the a u t o mated Senator to the i n p u t statements presented
him should simulate the r e p l i e s the r e a l Senator
might conceivably make.
Early t e s t s of such a
s i m u l a t i o n were not n o t a b l y s u c c e s s f u l ! but more
r e c e n t l y w i t h t h e i n c l u s i o n o f the i m p l i c a t i o n a l
molecule concept in the s i m u l a t i o n system f a f a i r
degree of success has been achieved. The d e t a i l s
of t h i s t e a t are to be r e p o r t e d in a separate
paper. How i m p l i c a t i o n a l molecules are used in
the s i m u l a t i o n program w i l l now be d e s c r i b e d .
There i s a l i s t B o f a l l s p e c i f i c b e l i e f
sentencea
in
the system. This l i s t is
Using I m p l i c a t i o n a l Molecules i n the Simulation
Tha h e a r t of the s i m u l a t i o n program is a subr o u t i n e which attempts t o " f i l l molecules", The
i n p u t is a s i n g l e sentence ( i n the simple S u b j e c t Verb-Object f o r m ) , and the output is e i t h e r a
f i l l e d molecule c o n t a i n i n g the i n p u t aentence and
the a p p r o p r i a t a o t h e r sentences needed to complete
the molecular s e t , o r a f a i l u r e s i g n a l i n d i c a t i n g
t h a t t h a system cannot e x p l i c a t e the i n p u t sentence.
F o l l o w i n g f a i l u r e t o e x p l i c a t e ! the s u b - r o u t i n e
may be entered again in an attempt to e x p l i c a t e the
opposite o f t h e i n p u t sentence t thus enabling the
syatem to d i s t i n g u i s h between unacceptable and
inexplicable inputs.
I f t h e m o l e c u l e - f i l l i n g r o u t i n e succeeds,
then of course t h e o t h e r sentences comprising the
completion o f t h e molecule provide the v e r b a l
materials f o r the e x p l i c a t i o n .
Interestingly,
the "name" of t h e molecule a l s o provides valuable
• x p l i c a t i v e m a t e r i a l f o r use i n e m b e l l i s h i n g the
There is a l i s t L of " l e g a l sentence t y p e s " ,
xaRpxceach s p e c i f y i n g a semantic a l l y and p r a g m a t i c a l l y reasonable s u b j e c t - v e r b - o b j e c t sequence.
Nonsensical combinations of categories, such as
inanimate subjects and animate v e r b s , do not
appear.
Nor do s u b j e c t i v e l y unbelievable t r i p l e t s —
f o r example, our simulated r i g h t - w i n g p o l i t i c i a n
does not a l l o w f o r the p o s s i b i l i t y of a Free-WorldNation A t t a c k i n g a Free-World-Nation, and t h i s sentence type does not appear on i t s l i s t L.
p a r t i t i o n e d i n t o blocks o f b e l i e f sentences o f
each given l e g a l sentence type ( f o r example, a l l
s p e c i f i c examples of Communist-Nations Subverting
Free-World-Nations).
This simple b l o c k i n g arrangement has been found to save a great amount of time
d u r i n g search and matching procedures using the
belief l i s t .
F i n a l l y , t h e r e i s a l i s t M o f molecules.
Each molecule is a p a r t i c u l a r s e t of s l e g a l
sentence types (most t y p i c a l l y , s=3) t o g e t h e r
w i t h a s p e c i f i c a t i o n of how p a r t i c u l a r elements are
to be l i n k e d across sentences. In g e n e r a l , the
molecular set Mh may be denotedt
with specific identifications of subscripts, for
example, i 1 = i 7 t h e index f o r the c a t e g o r y , A c t o r s ;
-643-
and t h i s fixedness i s t r a n s f e r r e d t o the a p p r o p r i ate element of another sentence type in t h e same
molecule.
Acceptable concrete sentences o f t h i s
type are generated from t h e a p p r o p r i a t e block of
the b e l i e f l i s t B , thus f i x i n g another f r e e
variable.
A t t h i s p o i n t , one remaining sentence
in t h e molecule t y p i c a l l y has b o t h X elements
f i x e d , and t h e value of t h e PTRUE f u n c t i o n is
calculated f o r i t .
Success e i t h e r completes t h e
molecule o r , i n case sentences remain i n t h e molecule s p e c i f i c a t i o n , new f r e e v a r i a b l e s are f i x e d
by concrete sentence g e n e r a t i o n and/or the PTRUE
f u n c t i o n i s c a l c u l a t e d f o r new f i x e d - v a r i a b l e
sentences•
The a b s t r a c t molecule Mh nay be s a i d to be
r a a l i a a d by a concrete
case
whenever a l l s e n tanoaa i n t h e a b s t r a c t molecule s p a c i f i c a t i o n ara
a p p r o p r i a t a l y e x e m p l i f i e d by sentencee which ara
preaumed t r u s . A santanca asaumas t h i s t r u t h
v a l u e , PTRUEt a i t h a r b y v i r t u a o f b a i n g tha i n p u t
t o t h a " f i l l - m o l e c u l e " r o u t i n e , o r bacausa i t i s
o n t h a b a l i a f l i s t , o r bacausa i t i s c a l c u l a t a d
t o b a a p p r o p r i a t a t o add t o t h a b a l i a f l i s t .
This
l a t t a r calculation involves a subtlaty to ba d i e cusaad s h o r t l y *
The PTRUE f u n c t i o n i n v o l v e s one or two s t e p s .
The f i r s t is a matching procedure a g a i n s t t h e
a p p r o p r i a t e block o f the B - l i s t , l o o k i n g f o r a
known sentence w i t h a l l elements " s i m i l a r " t o the
t e s t e d sentenoe.
( " S i m i l a r " is a more r e l a x e d
v e r s i o n o f " h i g h l y s i m i l a r " , mentioned p r e v i o u s l y . )
I f t h i s t e s t f a i l s , a second procedure i s undert a k e n , a procedure which lends t h s whole r o u t i n e
much g r e a t e r i n t e r e s t .
I n t h a process o f attamptad concrata r e a l i zations of a molecule, tha aavaral catagoriaa
X9R a c q u i r a a l a m a n t a l r e p r e s e n t a t i v e x , r .
Tha
c a t a g o r i a a may ba conaidarad " f r e e v a r i a b l e s " u n t i l
t h a program a l g o r i t h m assigns s p a c i f i c alamants t o
craata " f i x e d v a r i a b l e s " .
This t a r m i n o l o g y i a
naoaasary bacausa o f t h a l i n k a g e o f X - v a r i a b l e a
across sentences.
I t i s a convantion o f t h a p r o gram t h a t whenever t h e aama X-category appears in
d i f f e r e n t sentences, t h e elements r e p r e s e n t i n g these
c a t e g o r i e s should e i t h e r b e i d e n t i c a l o r h i g h l y
similar.
( " H i g h l y s i m i l a r " , i n t h a a p e c i a l senae
i n t e n d e d h e r e , means t h a t one element ia an i n stance o f t h e o t h e r .
These i n s t a n c e r e l a t i o n s h i p s
ara s t o r e d in the memory.
D e t a i l s have been given
in previous papers. 7,8,9)
This procedure i n t r o d u c e s r e c u r s i o n i n t h s
molecule-filling routine.
The second way in which
the f u n c t i o n PTRUE can be s a t i s f i e d f o r a sentence
i s b y e n t e r i n g t h a t sentence i n t u r n a s i n p u t t o
the f i l l i n g r o u t i n e .
A s u c c e s s f u l e x i t endows
the sentence w i t h presumptive t r u t h .
The e x i s t ence of t h i s r e a l i s t i c program device means t h a t
t h e attempt to complete molecules assumes an
open-ended c h a r a c t e r , w i t h p a r t i a l completions
l e a d i n g to new searches which may have p a r t i a l
c o m p l e t i o n s , and so o n . There is undoubtedly a
p s y c h o l o g i c a l l i m i t t o t h e depth such a r e c u r s i o n
might reasonably t a k e , and a p r e s e n t l y a r b i t r a r y
program parameter i s used t o d e f i n e t h i s l i m i t .
H o p e f u l l y , e m p i r i c a l work now i n progress w i l l
shed l i g h t upon t h e a p p r o p r i a t e value o f t h i s
parameter.
I f a l l f r e e v a r i a b l e s i n a molecule become
a p p r o p r i a t e l y f i x e d , t h e molecule i a s u c c e s s f u l l y
filled.
The c o n d i t i o n s of such success are appare n t l y q u i t e s t r a i g h t f o r w a r d mathematically.
Dej o t i n g t h e contents o f t h e concrete molecule
I n o r d e r t o make t h e f o r e g o i n g a b s t r a c t e x p l a n a t i o n of the r o u t i n e more c o n c r e t e , Table 2
gives some of the c a t e g o r i e s , elements, and molecules f o r t h e r i g h t - w i n g e r .
The components shown
o f course represent o n l y a s m a l l f r a c t i o n o f t h e
t o t a l s t o r a g e . Notice t h a t noun elements can bs
compound concepts such as s i t u a t i o n s - h e l p f u l - t o the-Communists.
One of the m o l e c u l e s , t h e c r u c i a l
one f o r present purposes, i s l a b e l e d "No-Win P o l i c y " by t h e system, and has t h e f o u r argumentst
L i b e r a l s , Western governments, s i t u a t i o n a - h e l p f u l to-the-Communists, and standing-up-to-the-Communists.
The sentences of t h e molecule a r e i Weetern
governments promote s i t u a t i o n s - h e l p f u l - t o - t h e Communists} Standing-up-to-the-Communiats prevents
situations-helpful-to-the-Communists1 Liberals
c o n t r o l Western governments; L i b e r a l s f e a r s t a n d i n g u p - t c - t h e - Communists•
Now suppose t h a t t h e i n p u t sentence is
" B r i t a i n n o t - i n t e r f e r e s - w i t h Czech i n v a s i o n . " The
sentence type is "Western governments promotes
s i t u a t i o n s - h e l p f u l - t o - t h e - C o m m u n i s t s . " The p r o gram c o n s u l t s i t s moleoule l i s t , and f i n d s t h a t t h s
No-Win P o l i c y molecule c o n t a i n s a sentenoe of t h i s
-644-
t y p e . Second molecule sentences are now generated
b y c o n s u l t i n g the b e l i e f l i s t t o l o c a t e sentences
of the t y p e , "standing-up-to-the-Communists p r e vanta s i t u a t i o n a - h e l p f u l - t o - t h e - C f c m m u n i s t s . " Not
a l l aantancaa o f t h i s typa are a p p r o p r i a t e , however , s i n c e the s p e c i f i c sentence o b j e c t muat be
s i m i l a r t o the f i x e d v a r i a b l e o f the f i r s t sentences
Czech i n v a s i o n s . Thus, " A l l i e d a i r l i f t
overcomes B e r l i n blockade" is not a p p r o p r i a t e . A
l i s t of candidate sentences is constructed, and we
suppose t h a t the f i r s t such sentence is "U.S. shows
Of force stops Red i n v a s i o n s . " For the t h i r d
sentence of the molecule, the r e q u i r e d sentence
type i s " L i b e r a l s c o n t r o l Western governments."
Here, the "Western governments" v a r i a b l e i s f i x e d
by the i n p u t as " B r i t a i n " , but " L i b e r a l s " is a
f r e e v a r i a b l e (because i t has not occurred b e f o r e ) .
Suppose "Wilson c o n t r o l s Britain 1 1 is the only approp r i a t e sentence o f t h i s t h i r d t y p e .
Passing f i n a l l y t o the f o u r t h sentence o f
the m o l e c u l e , we see t h a t the t y p e , " L i b e r a l s f e a r
standing-up-to-the-Communists" has been constrained
by f i x i n g the l i b e r a l as " W i l s o n " and standing-upto-the-Communists as " U . S . shows of f o r c e " or
something s i m i l a r . Now the program i n s t e a d of
g e n e r a t i n g p o s s i b i l i t i e s , must simply see i f anyt h i n g a t a l l o n the b e l i e f l i s t f i t s the s p e c i fication.
I f , s a y , "Wilson f e a r s U.S. nuclear
a c t i o n s " , i s p r e s e n t , then the molecule i s successf u l l y f i l l e d . The output can now be constructed
by c h a i n i n g t o g e t h e r the f o u r r e a l i z e d sentences of
t h i s m o l e c u l e , the name of the molecule, "No-Win
P o l i c y " and s u i t a b l e pre-keyed i n t e r s t i t i a l materi a l to produce a f l o r i d and smarmy account of why
B r i t a i n does not i n t e r f e r e w i t h the Czech i n v a s i o n .
I f , o n the o t h e r hand, the b e l i e f l i s t cont a i n a n o t h i n g to suggest t h a t "Wilson fears U.S.
shows o f f o r c e " ! then t h i s sentence i t s e l f i s i n put t o the f i l l - m o l e c u l e r o u t i n e .
Another molecule is found which w i l l accommodate a sentence of
t h i s t y p e i s a y , the " L i b e r a l Weakness" molecule —
" L i b e r a l s f e a r atanding-up-to-the-Communists",
" L i b e r a l s oppose armaments", "Armaments promote
standing-up-to-the-Communiets."
If the second and
t h i r d sentences here can be r e a l i z e d , say w i t h
"Wilson opposes B r i t i s h t r o o p commitments" and
" B r i t i a h t r o o p commitments a i d U.S. shows of f o r c e " ,
then "Wilson f e a r s U.S. shows of force"would be
considered presumptively t r u e by the program, thus
f i l l i n g the o r i g i n a l molecule.
F a i l u r e o f t h i a r e c u r s i o n leads t o a n i t e r a t i o n down the e a r l i e r l i s t o f candidate sentences
i n a n attempt t o f i n d another s u i t a b l e combination.
Exhaustion o f a l l candidate l i s t s leads t o abandonment of one molecule and attempts at f i l l i n g
o t h e r s , afterr which t h e f i n a l f a i l u r e e x i t occurs.
With a program l i m i t a t i o n on deep r e c u r s i o n ,
t h i s r o u t i n e u s u a l l y takes no more than h a l f a
minute. Of course t h i s sstimate depends on the
parameters of t h e p a r t i c u l a r memory, and we w i l l
need more experience to d e r i v e more general
estimates.
-645-
-647-