structure systems. - CiteSeerX

S e s s i o n 14 N a t u r a l Language: S e m a n t i c s and P a r s i n g CASE STRUCTURE SYSTEMS B e r t r a m Bruce Computer S c i e n c e Department Rutgers U n i v e r s i t y New B r u n s w i c k , New Jersey, U. S. A. Abstract I n s e c t i o n 3 a few r e p r e s e n t a t i v e systems w i t h cases a r e e x p r e s s e d i n t h e f o r m a l i s m s o f case l o g i c . S e c t i o n 4 c o n t a i n s a d e s c r i p t i o n of t h e CHRONOS 11 program3. T h i s program i s unusual i n h a v i n g a f l e x i b l e case s t r u c t u r e . For example i t has been u s e d w i t h one case s t r u c t u r e f o r t h e a n a l y s i s o f m e d i c a l r e c o r d s and w i t h a n o t h e r a s a model f o r glaucoma.

A l o g i c f o r c a s e s t r u c t u r e systems i s p r e s e n t e d w h i c h a l l o w s v a r i a t i o n s i n t h e o r d e r and number o f terms i n atomic f o r m u l a s . T h i s l o g i c i s used t o d e s c r i b e and c h a r a c t e r i z e f o u r e x i s t i n g case s y s t e m s . A c o m p u t e r p r o g r a m w h i c h a l l o w s f l e x i b l e case s t r u c tures is then described. A p p l i c a t i o n s o f t h e program t o m e d i c a l r e c o r d a n a l y s i s and disease m o d e l i n g are used t o i l l u s t r a t e i m p o r t a n t c o n c e p t s a b o u t case systems. S e v e r a l open p r o b l e m s a r e a l s o d i s c u s s e d . 1.

2.

Introduction

This section describes the e s s e n t i a l characteri s t i c s o f a case s t r u c t u r e system o r case l o g i c ( C L ) . I t a p p e a r s t h a t a n a d e q u a t e f o r m a l i z a t i o n can b e made w i t h i n a f i r s t order framework. The d i f f e r e n c e s between t h i s and a p u r e p r e d i c a t e calculus i n v o l v e t h e i n t r o d u c t i o n o f some f i x e d , o r c o n s t a n t p r e d i c a t e s , some a d d i t i o n a l a x i o m s , and some s y n t a c t i c a l changes t o r e f l e c t t h e f a c t t h a t c e r t a i n languages o f i n t e r e s t a l l o w v a r i a t i o n s i n t h e o r d e r and even o m i s s i o n s of some t e r m s .

Much o f t h e r e s e a r c h i n n a t u r a l l a n g u a g e p r o c e s s i n g has f o c u s e d o n t h e p r o b l e m o f s t o r a g e s t r u c t u r e s , o r t h e q u e s t i o n , "How i s t h e i n f o r m a t i o n or meaning of a s e n t e n c e to be r e p r e s e n t e d once a s e n t e n c e has been p a r s e d ? " Some o f t h e s t o r a g e s t r u c t u r e s have been s p e c i a l p u r p o s e o r g a n i z a t i o n s such as t h e SAD SAM p r o g r a m of L i n d s a y 12, t h e BASEBALL p r o g r a m of G r e e n , et al 8 , or the STUDENT program of Bobrow1. O t h e r s have been s e m a n t i c n e t works such a s TLC b y Q u i l l i a n 1 5 , PROTOSYNTHEX I I I by S i m m o n s 2 ^ , MEMOD by Norman and R u m e l h a r t 1 4 , or MENTAL b y S h a p i r o 1 9 , s t i l l o t h e r s have been based o n dependency grammar such a s S c h a n k ' s c o n c e p t u a l p a r ser-* 7 or on p r e d i c a t e c a l c u l u s , such as QA3 by Green and R a p h a e l 7 or PCDB by S a n d e w a l l 1 6 , A common f e a t u r e o f s e v e r a l o f t h e s e n a t u r a l l a n g u a g e p r o c e s s o r s i s a r e c o g n i t i o n o f deep case r e l a t i o n s a s d i s c u s s e d b y F i l l m o r e 6 . Some systems w h i c h have n o t e x p l i c i t l y used a case s t r u c t u r e have n e v e r t h e l e s s used c a s e - l i k e mechanisms. Jn a d d i t i o n , c a s e - l i k e systems have been used i n m o d e l i n g i n m e d i c i n e and p s y c h o l o g y , even w i t h o u t n a t u r a l l a n g u a g e . I t a p p e a r s now t h a t a f o r m a l i z a t i o n o f case s t r u c t u r e l a n g u a g e s w o u l d b e u s e f u l and f e a s i b l e . On t h e one hand t h e l a r g e amount o f d a t a o n case systems a p p e a r s t o b e s t r u c t u r e d i n s i m i l a r ways and y e t d i f f e r e n c e s i n t e r m i n o l o g y and a p p r o a c h have c r e a t e d serious communication problems. On the other hand, even i f w e a d m i t t h a t t h e u s u a l f i r s t o r d e r l o g i c s somehow c a p t u r e t h e e s s e n t i a l n a t u r e o f t h e s e s y s t e m s , t h e s i m p l i f i c a t i o n s imposed t h e r e f o r e l e g a n c e and t h e o r e t i c a l u s e f u l n e s s make them l e s s u s e f u l as a means f o r e x a m i n i n g and c o m p a r i n g case

structure

Case L o g i c

systems.

This p a p e r p r e s e n t s ( i n s e c t i o n 2 ) a l a n g u a g e f o r case s t r u c t u r e systems w h i c h i s b u i l t o n f i r s t o r d e r l o g i c and y e t a l l o w s such t h i n g s a s o m i t t e d and rearranged terms i n atomic f o r m u l a s , p r e p o s i t i o n s , and t h e c h e c k i n g o f case r e s t r i c t i o n s i n t h e d e t e r mination of well-formedness. The case l o g i c s h o u l d make i t s i g n i f i c a n t l y e a s i e r t o f o r m u l a t e , a n a l y z e o r compare case s t r u c t u r e s y s t e m s .

this r e s e a r c h , was s u p p o r t e d by t h e N a t i o n a l I n s t i t u t e s o f H e a l t h . Resource o n Computers i n B i o m e d i c i n e a t R u t g e r s U n i v e r s i t y ( G r a n t R R 643)

364

The a x i o m system f o r a case l o g i c depends h e a v i l y o n t h e s p e c i f i c cases w h i c h a r e used a n d , o f course, on the i n t e r p r e t a t i o n s we wish to a l l o w . Because n o axioms a r e g i v e n h e r e i t i s p e r h a p s more appropriate to c a l l CL a " n o t a t i o n a l system" r a t h e r than a " l o g i c " . However, i t s h o u l d b e remembered t h a t in any p a r t i c u l a r CL one must i n c l u d e an axiom system a s w e l l a s naming t h e v a r i o u s s y n t a c t i c and s e m a n t i c o b j e c t s i n t h e s y s t e m .

order

W e have a l l o f t h e u s u a l logic:

(p =

{p(ni)}

is

s

symbols f o u n d i n f i r s t

countable set of n . - a r y p r e d i c a t e s ,

365

Then the set of atomic formulas is simply the set of those kernel s t r i n g s which have n terms f o r an n-ary p r e d i c a t e , a l l in the usual order. The semantic s t r u c t u r e s and i n t e r p r e t a t i o n s are unchanged, but the denotation of an n-ary p r e d i c a t e must n e c e s s a r i l y be an n-ary r e l a t i o n and the case sequence i s , of course, irrelevant. This s e c t i o n presents a n o t a t i o n f o r a class of systems c a l l e d "case l o g i c s " . I t i s reasonable t o ask how the case systems of l i n g u i s t s , p s y c h o l o g i s t s , and computer s c i e n t i s t s f i t i n t o t h i s framework and p a r t i c u l a r l y how the d i s t i n c t i o n between surface and deep case systems is handled. It is o f t e n argued t h a t a language has at l e a s t two case systems, one at the surface or s y n t a c t i c l e v e l and the other at a deep or conceptual l e v e l . The n o t a t i o n presented here is designed f o r both l e v e l s . Of course, at the deep l e v e l , we no longer have p r e p o s i t i o n s in the usual sense, but we do have s t r u c t u r e s which can be considered to be atomic formulas of CL , w i t h v a r i a b i l i t y in the order and presence of terms. Part of the understanding process is then a t r a n s l a t i o n from one case system to another. Usually a deep case system has very few cases, each one j u s t i f i e d in terms of i t s presence is basic semantic c o n s t r u c t s . A surface system may have e i t h e r a p r o f u s i o n of cases corresponding to the many d i f f e r ent s y n t a c t i c p o s s i b i l i t i e s or a small set of r a t h e r amorphous cases, such as " o b j e c t i v e " . The t r a n s l a t i o n from surface to deep cases o f t e n involves an expansion in storage s t r u c t u r e s wherein a s i n g l e p r e d i c a t e w i t h a few loosely defined arguments becomes a complex of more p r i m i t i v e predicates w i t h arguments o f t e n missing but always w e l l defined conceptually. I t i s the nature o f the missing arguments which p o i n t s out connections to other p a r t s of the discourse. I n special s i t u a t i o n s t h i s complex o f p r i m i t i v e predicates can be s i m p l i f i e d by e l a b o r a t i n g the surface case system (see s e c t i o n 4 ) . P o s i t i o n s of nodes in the semantic or conceptual network are in e f f e c t named by new surface cases whose case condit i o n s check to see if a term could f i l l t h e approp r i a t e conceptual p o s i t i o n . The r e s u l t is a loss in i n f o r m a t i o n w i t h a corresponding improvement in e f f i c i e n c y f o r the special s i t u a t i o n . I t appears t h a t t r e a t i n g such t h i n g s as causation and purpose as cases ( r a t h e r than as second order predicates on sentences) is such a s i m p l i f i c a t i o n . The next s e c t i o n discusses some w e l l known case systems in terms of CL. While space l i m i t a t i o n s make the d e s c r i p t i o n s r a t h e r spare, they should i n d i c a t e at l e a s t how some of the e s s e n t i a l p a r t s of the systems would be expressed. The problems of surface v s . deep and f i r s t order v s . second order f o r any p a r t i c u l a r system would need to be discussed at much greater l e n g t h . 366

3. Some Representative Case Structure Systems

3.2 Simmons -i Semantic Networks

In t h i s s e c t i o n we examine a v a r i e t y of systems which appear s u i t a b l e f o r d e s c r i p t i o n as case s t r u c t u r e systems. For the sake of c l a r i t y , the special case C o , which is always assumed to be in C is not

R.F. Simmons20 gives an extensive account of semantic networks and t h e i r use in processing n a t u r a l language by computer. An e s s e n t i a l f e a t u r e of these networks is a set of deep case r e l a t i o n s connecting nominal concepts (underlying noun phrases) to verbs. While a semantic network does not r e q u i r e any s p e c i f i c case s t r u c t u r e Simmons has chosen one which follows from work of Celce-Murcia and F i l l m o r e " .

l i s t e d i n the d e s c r i p t i o n s . 3.1

F i l l m o r e - Natural Language Cases

The most persuasive argument f o r the u n i v e r s a l i t y of cases in the deep s t r u c t u r e s of n a t u r a l languages is made by F i l l m o r e 6 . He argues from evidence in many languages t h a t every language, regardless of i t s surface s t r u c t u r e s , has a deep case system which gives the r e l a t i o n s h i p of many nouns to a verb.

In h i s system there are f i v e deep case r e l a t i o n s , so t h a t C = ( causal actant, thome, l o c u s , source, and g o a l } . The set of p r e p o s i t i o n s , or Q., is of course j u s t the p r e p o s i t i o n s of English. Verbs are c l a s s i f i e d i n t o paradigms according to the case sequences they allow, a paradigm, t h e n , is a case s t r u c t u r e ( 3 ). For example, the ergative paradigm consists of the sequences ( f o r the a c t i v e v o i c e ) :

F i l l m o r e ' s d e s c r i p t i o n of cases and t h e i r purpose in language can be expressed d i r e c t l y in CL. He suggests a minimum of s i x cases.

( c a u s a l - a c t a n t . , theme, causal-actant ) (causal a c t a n t . , theme)

Agentive (A) - the animate perceived i n s t i g a t o r of an action Instrumental [ I ) - the inanimate force or object which is causally i n v o l v e d . Dative (D) - the animate being a f f e c t e d . F a c t i t i v e (F) - the object or being r e s u l t i n g from the a c t i o n . Locative (L) - l o c a t i o n or s p a t i a l o r i e n t a t i o n . Objective (0) - determined by verb (neutral case).

(causal a c t a n t - , theme) (theme) Simmons gives "break" as an example of an ergative verb. Thus John broke John broke The hammer The window

While other possible cases are mentioned we can say roughly t h a t C = {A, I, D, F, L, 0 } . F i l l m o r e notes t h a t the cases A and D r e q u i r e animate nouns, t h a t i s , f o r a noun to serve in the agentive or d a t i v e case it must have the f e a t u r e "animate". Other features making up the set m can e a s i l y be imagined f o r the other cases.

the window with a hammer. the window broke the window broke

are a l l well-formed since in each case one of the case sequences is matched (where "John" is the causala c t a n t . , "window" is the theme, and "hammer" is the causal-actant

).

Another example is the " r e f l e x i v e - d e l e t i o n paradigm where the theme is deleted if it corresponds to the CA1 [ c a u s a l - a c t a n t . ] " . Thus " r u n " may be used

Corresponding to features on nouns Fillmore says t h a t verbs can be c l a s s i f i e d according to "sentence types" or "case frames". A case frame t e l l s what case r e l a t i o n s h i p s may e x i s t between a verb and i t s nouns. For example "open" may be used in four ways:

in several ways: John ran to school John ran a machine The machine ran The brook ran

The door opened. (0) John opened the door. (A 0) The wind opened the door. (I 0) John opened the door w i t h a c h i s e l , (A 0 I )

In each of the sentences there is a theme - John, machine, or brook. The paradigm allows the d e l e t i o n of the theme if it is the same as the causal-actant. In t h i s the case s t r u c t u r e is

We can represent the case frame f o r "open" as [0 [A) ( I ) ] . This says t h a t when "open" is used i t s object must appear but the agent and instrument are o p t i o n a l . C l e a r l y a case frame is a case s t r u c t u r e

(causal-actant, goal) (causal-actant, theme) (theme)

(.£ ) in CL.

F i l l m o r e shows by example the case signals (S.) of various languages. He also gives some t e n t a t i v e r u l e s f o r English. For example:

Simmons discusses other aspects of the case s t r u c t u r e p o r t i o n of semantic networks as w e l l . He p o i n t s out t h a t a "semantic d e f i n i t i o n " of a verb can be given by d e s c r i b i n g the p r o p e r t i e s of the nominal concepts serving in each case r e l a t i o n s h i p to the verb. It appears to be the case t h a t these d e s c r i p t i o n s can be adequately handled by means of our case conditions. If so then a parser f o r semantic networks would simply check the d e s c r i p t i o n s in the process of determining well-formedness.

"The A p r e p o s i t i o n is by; the I p r e p o s i t i o n is by if there is no A, otherwise it is w i t h ; the 0 and F p r e p o s i t i o n s are t y p i c a l l y zero; the B [benefactive case] p r e p o s i t i o n is f o r ; the D p r e p o s i t i o n is t y p i cally t o . . . . " I f t h e r e is an A it becomes the s u b j e c t ; otherwise, i f there i s a n I , i t becomes the s u b j e c t ; otherwise the subject i s the 0 . "

367

368

For example the noun human might have a property " s t a t e - o f - m i n d " w i t h allowed values being "happy", " s a d " , or " b l u e " . It could have the superset " a n i m a l " and the subset "woman", N a t u r a l l y any noun w i l l have the p r o p e r t i e s o f i t s supersets i n a d d i t i o n t o those e x p l i c i t l y l i s t e d . A verb entry has the form: [word]

The paradigm f o r a verb is an ordered set of cases. In a given simple sentence CHRONOS II w i l l ( c u r r e n t l y ) allow any or a l l of the cases to be missing. Thus the case s t r u c t u r e f o r a verb is a p o r t i o n of the power set of paradigm subject to r e s t r i c t i o n s on features. Case determination depends on s y n t a c t i c a l i n f o r mation (case s i g n a l s ) as w e l l as feature checking (case c o n d i t i o n s ) . Currently in CHRONOS II the def i n i t i o n s of case signals and case conditions are combined in one LISP f u n c t i o n f o r each case. Someone who wishes to use CHRONOS II f o r a s p e c i f i c a p p l i c a t i o n must specify the l i s t of cases, define each case, and modify the lexicon a p p r o p r i a t e l y . Each possible case f u n c t i o n is c a l l e d f o r every p r e p o s i t i o n term (noun phrase, p r e p o s i t i o n a l phrase, or adverb) and returns a number g i v i n g the l i k e l i h o o d t h a t the given p r e p o s i t i o n term serves t h a t case r e l a t i o n s h i p to the main verb of the sentence. The case whose f u n c t i o n r e t u r n s the highest value is temp o r a r i l y assigned to the p r e p o s i t i o n term. A f a i l u r e , c o n s i s t i n g of values of D f o r a l l the cases, forces a backup to the previous term and a reassignment of the case f o r t h a t term. The case parser 2 1 is e s s e n t i a l l y independent of the p a r t i c u l a r set of cases being used. I t s speed and g e n e r a l i t y are of course d i r e c t l y dependent on the accuracy of the l i k e l i h o o d numbers returned by the case f u n c t i o n s . There are various features which extend the simple h e u r i s t i c value assignments, making the case functions easier to w r i t e or the parsing more e f f i c i e n t . For example, a v a r i a b l e number of "pre-emptive l e v e l s " can be created. Once a case f u n c t i o n r e t u r n s a value at a new and higher preemptive l e v e l then only values at t h a t l e v e l (or higher) are considered v a l i d . This is e s p e c i a l l y usef u l when one knows t h a t a p a r t i c u l a r p r e p o s i t i o n , say, signals one and only one case. Then t h a t case funct i o n r e t u r n s a value which pre-empts any previous use of t h a t case. CHRONOS II has been used p r i m a r i l y w i t h two case systems. One is f o r the analysis of medical records, i . e . nurse's n o t e s , and is the more general of the two. I n i t c = { a g e n t i v e , o b j e c t i v e , d a t i v e , locat i v e , i n s t r u m e n t a l , temporal, frequency, d e r i v a t i v e , manner) where a g e n t i v e , o b j e c t i v e , d a t i v e , i n s t r u mental, and l o c a t i v e are as in F i l l m o r e 1 0 , temporal points to the d u r a t i o n of an event, frequency is i n d i c a t e d by adverbs such as " r a r e l y " and " o f t e n " , d e r i v a t i v e is i n d i c a t e d by words such as " i n c r e a s i n g " , and manner includes most other adverbs. The e f f e c t of such a case system can best be seen by comparison with a t r a d i t i o n a l C = {subject, object, indirect o b j e c t } system. For example, the parse of "John

369

s t r u c t u r e transformations.

Note t h a t in the second parse, "pressure" has become the p r e d i c a t e . It is an essential p r i n c i p l e of CHRONOS II t h a t the p r e d i c a t e of a simple sentence is the l o g i c a l p r e d i c a t e ( r e l a t i v e t o the purpose o f d i s cussion) r a t h e r than the s y n t a c t i c a l p r e d i c a t e (usually the v e r b ) . 5.

Ac knowledgements Many people have c o n t r i b u t e d to t h i s work. Discussions w i t h Robert Simmons f i r s t i n t e r e s t e d me in case l o g i c s . W i l l i a m Fabens, Richard Orgass, and Ann Yasuhara made u s e f u l suggestions on case l o g i c , N e i l Singer, Santosh Chokhani, Nanette Yaskin, and Mike T r i g o b o f f a l l worked on CHRONOS II and cont r i b u t e d to discussions on case systems. I also want to thank the referees f o r t h e i r comments and suggestions which I have i n c o r p o r a t e d .

Discussion

This r e p o r t presents a n o t a t i o n a l system (CL) f o r case s t r u c t u r e systems and examines a few e x i s t i n g systems in terms of the n o t a t i o n . CL is a f i r s t order system which appears u s e f u l in c h a r a c t e r i z i n g case systems and w i l l h o p e f u l l y lead to a b e t t e r understanding of n a t u r a l language storage s t r u c t u r e s .

References

Several observations about case s t r u c t u r e systems f o l l o w from t h i s s t u d y , observations which are a c t u a l l y open problems. For example, consider the sentence "John gave b i r t h to a baby". The oddness of t h i s sentence can be handled in at l e a s t three ways. F i r s t , we could say t h a t " t o give b i r t h " is a p r e d i cate which imposes the case c o n d i t i o n on the agent t h a t it be female. In t h a t case the sentence is not a formula of CL. CHRONOS II could do t h a t , or a l t e r n a t i v e l y , r a i s e i t s eyebrows by accepting the sentence at a lower pre-emptive l e v e l (see section 4 ) . Second, we could say t h a t " t o g i v e " is a p r e d i c a t e , t h a t the sentence i s w e l l - f o r m e d , but i t s negation i s provable from a set of u s u a l l y accepted f a c t s about l i f e . T h i r d , we could say t h a t " t o g i v e " is the p r e d i c a t e but demand t h a t the sentence be treated as i l l - f o r m e d . The l a s t approach is f e a s i b l e as long as the condit i o n s on well-formedness can be stated as predicates on the s i n g l e sentence (such as features) but is probably r i s k y i f i t requires examination o f other formulas. An i n t e r e s t i n g problem, proposed in s e c t i o n 2, concerns the i n t e r a c t i o n of case signals and case conditions. In CHRONOS II these functions mesh in a r a t h e r haphazard way. It should be i n t e r e s t i n g to examine s i t u a t i o n s where p r i o r i t i e s in t h e i r use affect efficiency. Case systems, since they emphasize a l o g i c a l s t r u c t u r e , r a t h e r than a p u r e l y s y n t a c t i c one, lessen the importance of the s y n t a c t i c predicate and of the sentence as a u n i t . Schank 17 and others have stressed the need to examine a discourse as an i n t e g r a t e d whole r a t h e r as a c o l l e c t i o n of i s o l a t e d sentences. It seems t h a t the b e t t e r a case system i s , t h a t is, the more r e l e v a n t i t i s t o the problem s o l v i n g s i t u a t i o n at hand, the easier it is to connect sentences in the discourse in meaningful ways. Since many case systems are in use, a n a t u r a l question a r i s e s , "What is the best case system?" My experience w i t h CHRONOS II suggests t h a t , at t h i s stage in the development of i n t e l l i g e n t programs, one can only speak of the goodness of a case system r e l a t i v e to a problem s i t u a t i o n . The f i n e l y d i s t i n guished cases which are a r i s i n g in the medical model v e r s i o n of CHRONOS II would probably only c l u t t e r the program which analyzes n u r s e ' s notes. Conversely, c e r t a i n cases which one would need in discussing everyday l i f e would be unnecessary in a medical model. An important problem then is to decide what cases to use in a p a r t i c u l a r a p p l i c a t i o n . Even more i n t e r e s t i n g might be a study of transformations between various case r e p r e s e n t a t i o n s . Problems of summariz a t i o n and a n a l o g i c a l reasoning w i l l probably be more t r a c t a b l e w i t h a b e t t e r understanding of case

370

1.

Bobrow, D a n i e l . " N a t u r a l language i n p u t f o r a computer problem s o l v i n g system". In Semantic Information Processing (Ed: Marvin Minsky) MIT Press, 1968.

2.

Bruce, Bertram. "A model f o r temporal references and i t s a p p l i c a t i o n in a question answering program". A r t i f i c i a l I n t e l l i g e n c e 3, 1972 pp. 1-26.

3.

Bruce, Bertram, and N e i l Singer, "CHRONOS I I : the processing of incompletely s p e c i f i e d d a t a " . Computer Science Department, Rutgers-NIH r e p o r t CBM-TR-9, 1972.

4.

Celce-Murcia, M. English Comparatives. Thesis, UCLA, 1972.

5.

Chomsky, Noam. Aspects of the Theory of Syntax. MIT Press, 1964.

6.

F i l l m o r e , Charles. "The case f o r case". In Universals in L i n g u i s t i c Theory (Ed; Bach and Harms), H o l t , Rinehart, and Winston, 1968.

7.

Green, C o r d e l l , and B. Raphael. "Research on i n t e l l i g e n t question answering systems". Proc. ACM 23rd Nat. Conf., 1968.

8.

Green, B . F . , A.K. Wolf, Carol Chomsky, & Kenneth Laughery. " B a s e b a l l : An Automatic Question Answerer". In Computers and Thought (Ed: Feigenbaum and Feldman), McGraw-Hill, 1963.

9.

Kaplan, R.M. "Augmented t r a n s i t i o n networks as psychological models of sentence comprehension". A r t i f i c i a l I n t e l l i g e n c e 3, 1972. pp. 77-100.

Ph.D.

10.

Katz, J . J . The Philosophy of Language. 1965.

11.

K u l i k o w s k i , C. and S. Weiss. "Computer based models f o r glaucoma". Computer Science Department, Rutgers-NIH r e p o r t CBM-TR-3, 1971.

12.

Lindsay, Robert K. " I n f e r e n t i a l memory as the basis of machines which understand language". In Computers and Thought (Ed: Feigenbaum and Feldman), McGraw-Hill 1963.

13.

M a r t i n , W i l l i a m A . , Rand B. Krumland, and Alex. Sunguroff. "More MAPL: S p e c i f i c a t i o n s and basic s t r u c t u r e s " . Automatic Programming Group, Project Mac, MIT, I n t e r n a l Memo 8, 1973.

14.

Norman, Donald A. and David E. Ruraelhart. " A c t i v e semantic networks as a model of human memory". Third I n t e r n a t i o n a l J o i n t Conf. on A r t i f i c i a l I n t e l l i g e n c e , 1973

15.

Q u i l l i a n , Ross. "The teachable language comprehender: a simulation program and theory of language". Comm. ACM 12, 1969, pp. 459-476.

16.

Sandewall, E r i k . "A programming t o o l for management of a p r e d i c a t e - c a l c u l u s - o r i e n t e d data base". Second I n t ' l .Joint..Conf\. on A r t i f i c i a l I n t e l l i g e n c e , 1971.

17.

Schank,. Roger. "Finding the conceptual content and i n t e n t i o n in an utterance in natural language conversation". Second J n t ' l J o i n t Conf. on A r t i f i c i a l I n t e l l i g e n c e , 1971.

18.

Schwarcz, R.M., J . F , Burger, 6 R.F. Simmons. "A deductive question answerer f o r natural language i n f e r e n c e " . Comm. ACM 13, 1970, pp. 167-163.

19.

Shapiro, S t u a r t . "A net s t r u c t u r e f o r semantic i n f o r m a t i o n storage, deduction and r e t r i e v a l " . Second I n t ' 1 J o i n t Conf. on A r t i f i c i a l I n t e l l i g e n c e , 1971.

20.

Simmons, R.F. "Semantic networks: t h e i r comp u t a t i o n and use f o r understanding English sentences". Computer Science Department U n i v e r s i t y of Texas at Austin-CAI Lab r e p o r t NL-6, 1972.

21

T r i g o b o f f , Mike. "A h e u r i s t i c case s t r u c t u r e p a r s e r " . Computer Science Department, RutgersNIH r e p o r t CBM-TM-16, 1973.

22.

Woods, W.A. " T r a n s i t i o n network grammars f o r n a t u r a l language a n a l y s i s " . Comm. ACM 13, 1970, pp. 591-602.

371