systems are semantic driven and they do only local syntactic checks during analysis.. In designing SHEILA we chose another way of integrating syntax with semantics.. The aim of syntactic
Trang 1T
INTEGRATING SEMANTICS kNO FLEXIBLE SYNTAX BY EXPLOITING ISONORPHISM BETWEEN GRAIelATICAL AND SEMANTICAL RELATIONS
N o r e n a D a n i e l i , F r a n c o F e r r a r a , R o b e r t o G e m e l l o , C 1 a u d i o R u l l e n t CSELT - C e n t r o S t u d i • L a b o r a t o r i T e l e c o a u n i c e z l o n i -
V i a G R e l s s R o a o l i 2 7 4 , 1 0 1 4 8 T o r i n o , ITALY
• ABSTRACT
T h i s w o r k c o n c e r n s i n t e g r a t i o n between s y n t a x
and s e m a n t i c s S y n t a c t i c and s e m a n t i c a c t i v i t i e s
r e l y on s e p a r a t e b o d i e s o f k n o w l e d g e s I n t e g r a t i o n
i s o b t a i n e d by e x p l o i t i n g t h e isomorphism between
g r a m m a t i c a l r e l a t i o n s (among i m m e d i a t e c o n s t i t u ~
e r r s ) and c o n c e p t u a l r e l a t i o n s , t h a n k s t o a l i m i t e d
s e t o f f o r m a l mapping r u l e s S y n t a c t i c a n a l y s i s
does n o t c o n s t r u c t a l l t h e e x p l i c i t p a r s e t r e e s b u t
j u s t a g r a p h t h a t r e p r e s e n t s a l l t h e p l a u s i b l e
g r a m m a t i c a l r e l a t i o n s among i m m e d i a t e c o n s t i t u e n t s
Such graph g i v e s t h e s e m a n t i c i n t e r p r e t e r , based on
C o n c e p t u a l Graphs f o r m a l i s m , t h e d i s c r i m i n a t i v e
power r e q u i r e d t o e s t a b l i s h c o n c e p t u a l r e l a t i o n s
I INTRODUCTION
I n t h e f i e l d o f a u t o m a t i c n a t u r a l l a n g u a g e
u n d e r s t a n d i n g , t h e p r o b l e m o f c o n n e c t i n g s y n t a x and
s e m a n t i c s has been f a c e d i n t h r e e d i f f e r e n t ways
Some a u t h o r s a r e persuaded t h a t u n d e r s t a n d i n g
n a t u r a l l a n g u a g e r e q u i r e s no use o f s y n t a c t i c know-
l e d g e They c l a i m t h a t s e m a n t i c r e p r e s e n t a t i o n can
be b u i l t d i r e c t l y from t h e s u r f a c e s t r i n g , w i t h o u t
t h e h e l p o f a l m o s t any s y n t a c t i c s o u r c e ( 1 )
Other a u t h o r s p r o p o s e d h i g h l y s y n t a c t i c s y s -
tems, s t a r t i n g from t h e i d e a t h a t t h e r e p r e s e n -
t a t i o n o f t h e s y n t a c t i c s t r u c t u r e i s p r e l i m i n a r y t o
t h e u n d e r s t a n d i n g p r o c e s s ( 2 )
While the work of t h i s second group of resear-
chers was concerned mainly with the understanding
of i n d i v i d u a l sentences, the work of the partisans
of semantics was about the understanding of whole
t e x t s
T h i s s h i f t i n g o f a t t e n t i o n s u b s t a i n e d t h e i d e a
t h a t s y n t a x and s e m a n t i c s s h o u l d be used i n an
i n t e g r a t e d way Most r e s e a r c h e r s have t h o u g h t t h a t
s e m a n t i c s and s y n t a x s h o u l d be i n t e g r a t e d w i t h
r e s p e c t t o b o t h t h e r e p r e s e n t a t i o n and t h e p r o -
c e s s i n g ( 3 ) ; o t h e r s have c l a i m e d t h a t i t i s more
e f f i c i e n t t o b u i l d a f u l l - b l o o d e d s y n t a c t i c r e p r e -
s e n t a t i o n d u r i n g t h e p a r s i n g p r o c e s s ( 4 )
(1) See t h e system IPP [Schank 8 0 ]
(2) The LUNAR system [Woods ?2] i s a c l a s s i c a l
example
(3) An example i s t h e C o n c e p t u a l A n a l y z e r [Birnbaum
8 1 ]
(4) See MOPTRANS [ L y t i n e n 8 5 ] ,
Our approach s h a r e s some c o m m u n a l i t i e s w i t h t h e
l a s t p o s i t i o n We r e c k o n t h a t s e m a n t i c and s y n t a c -
t i c p r o c e s s e s s h o u l d r e l y on s e p a r a t e k n o w l e d g e
b o d i e s Uur e f f o r t i s m a i n l y f o c u s e d on t h e r e a l i -
z a t i o n o f t h e i n t e g r a t i o n by e x p l o i t i n g t h e i s o - morphism between s y n t a c t i c s t r u c t u r e s and s e m a n t i c
r e p r e s e n t a t i o n s , r a t h e r t h a n by making s y n t a c t i c and s e m a n t i c p r o c e s s e s i n t e r a c t , as i t happens i n
p r e v i o u s i n t e g r a t e d p a r s e r s ( 5 ) The i d e a o f i s o - morphism i s n o t c a r r i e d o u t t h r o u g h o n e - t o - o n e c o r - respondence between s y n t a c t i c r u l e s and s e m a n t i c ones - as i n M o n t a g u e - i n s p i r e d p a r s e r s ( 6 ) , b u t by mapping i n a f o r m a l way g r a m m a t i c a l and c o n c e p t u a l
r e l a t i o n s The use o f g r a m m a t i c a l r e l a t i o n s as i n -
t e r m e d i a t e l e v e l between s y n t a x and s e m a n t i c s was
a l s o a d o p t e d i n t h e KING KONG p a r s e r ( 7 ) , b u t t h i s system i s s t i l l more n e a r t o t h e p o s i t i o n which wants t h e r e p r e s e n t a t i o n o f s y n t a x and s e m a n t i c s as
w e l l as t h e i r p r o c e s s e s t o i n t e r a c t , w h i l e o u r
c h o i c e i s t o m a i n t a i n s e p a r a t e t h e s e d i f f e r e n t sources o f k n o w l e d g e
The subsequent p a r a g r a p h s d e s c r i b e how t h i s
h y p o t h e s i s works i n S H E Z L A ( S y n t a x H e l p i n g
E x p e c t a t i o n s In Language A n a l y s i s ) , a p r o t o t y p e
d e v e l o p e d a t CSELT l a b o r a t o r i e s ( T u r i n , I t a l y ) The aim o f SHEILA i s t o a n a l y z e and t o e x t r a c t r e l e v a n t information from news (coming from the I t a l i a n news agency "ANSA") The system is i n i t i a l l y being applied to t e x t s describing v a r i a t i o n s in the top- management of commercial s o c i e t i e s ; i t has been
f u l l y implemented on a Symbolics Lisp machine SHEILA takes advantage both from the use of expec-
t a t i o n s and from the combination of the r e s u l t s of
a non-conventional s y n t a c t i c analysis with the
a c t i v i t y of a surface semantic analysis, based on the formalism of conceptual graphs (8) In t h i s paper we describe j u s t the p r i n c i p l e s which guide the i n t e g r a t i o n between syntax and semantics SHEILA c o r r e c t l y analyzes a set of t h i r t y news, generating f o r each of them a set of records f o r a
r e l a t i o n a l data base
2 THE PROBLEM AND OUR PROPOSAL
I n t e x t u n d e r s t a n d i n g systems s y n t a x and seman-
t i c s have a l m o s t a l w a y s been d e a l t w i t h i n t e g r a -
t i o n o f t h e i r p r o c e s s i n g U s u a l l y t h i s k i n d o f (5) See PSLI3 [ F r e d e r k i n g 8 5 ] , FZDO [Lesmo 85] and WEDNESDAY-2 [ S t o c k 8 6 ]
(6) See ABSITY [ H i r s t 8 4 ] (7) See [BAYER 65]
(8) See [Sowa 84] and a l s o t h e f o u r t h p a r a g r a p h b e - low
Trang 2systems are semantic driven and they do only local
syntactic checks during analysis Doing local syn-
t a c t i c checks only involves l i t t l e amount of syn-
t a c t i c knowledge and that is misleading in solving
problems as anaphoric reference, prepositional
attachment, conjunction and so on
In a d i f f e r e n t approach t h e i n t e g r a t i o n has
been r e a l i z e d d u r i n g t h e s y n t a c t i c s t r u c t u r e r e p r e -
s e n t a t i o n construction: the syntactic parser makes
use of semantic information to handle s t r u c t u r a l
ambiguities
The q u e s t i o n i n g done by t h e s y n t a c t i c a n a l y z e r
t o t h e s e m a n t i c component a~ms t o c u t down t h e
number o f p a r s e t r e e s , b u t v e r y many r u l e s a r e
r e q u i r e d f o r t h i s q u e s t i o n i n g , which has a l w a y s
been t h e most d o m a i n - d e p e n d e n t p a r t o f n a t u r a l
language u n d e r s t a n d i n g systems
In designing SHEILA we chose another way of
integrating syntax with semantics The basic schema
may l o o k r a t h e r c l a s s i c : t h e system produces a s y n -
t a c t i c a n a l y s i s o f t h e t e x t , d r i v e n on t h e b a s i s o f
p u r e l y s y n t a c t i c k n o w l e d g e The s e m a n t i c a n a l y z e r
checks t h e s y n t a c t i c o u t p u t t o see i f t h e s e m a n t i c
r e l a t i o n s among words a r e s u p p o r t e d by i t
But a classical s y n t a x - f i r s t analysis is highly
i n e f f i c i e n t I t cannot solve s t r u c t u r a l ambiguities
without the help of any semantic source and that
leads to an explosion of the number of syntactic
parse trees, some of them representing a r t i f i c i a l
syntactic ambiguities So there are two problems:
reducing the explosion of ambiguities and deter-
mining how semantic patterns for each word i n t e r a c t
with s y n t a x
Our proposal faces these problems through the
o r i g i n a l combination of two key ideas, i e :
I) a f l e x i b l e syntactic analysis, which is per-
formed by constructing not a l l the e x p l i c i t
parse trees but j u s t a graph, representing a l l
the plausible grammatical r e l a t i o n s among imme-
diate constituents;
2) a formal way of i n t e r a c t i o n between syntax and
semantics e x p l o i t i n g the isomorphism between
s y n t a c t i c s t r u c t u r e s ( g r a m m a t i c a l r e l a t i o n s
among i m m e d i a t e c o n s t i t u e n t s ) and s e m a n t i c ones
( c o n c e p t u a l r e l a t i o n s )
Such f l e x i b l e syntactic analysis gains a
discriminative power ( s u f f i c i e n t for aiding seman-
t i c s in solving ambiguities) and avoids the explo-
sion in the parse trees number Furthermore, the
mapping between grammatical and conceptual r e l a -
tions can be defined through a limited set of f o r -
mal rules
3 THE SYNTACTIC ANALYSIS
Our system has the goal of generating a seman-
t i c structure that has to be consistent with the
syntactic form used to convey i t in the t e x t The
aim of syntactic analysis is to support semantics
A f i r s t a c t i v i t y performed by the syntactic
@nalyzer is the recognition of constituents of the
phrase structure of t e x t This is done by applying
a set of r e w r i t i n g phrase structure rules f o r
I t a l i a n language These rules u t i l i z e the output of
a p r e v i o u s m o r p h o l o g i c a l a n a l y s i s t h a t a s s i g n s t o words m o r p h o l o g i c a l and l e x i c a l f e a t u r e s ( g e n d e r , number, l e x i c a l c a t e g o r y and so o n )
In t h i s analysis phase the application of the syntactic rules is limited to the recognition of the basic constituents of the phrase structure of the sentences A basic constituent (BC, henceforth)
is a NP, a PP or a VP described at a minimal level
of complexity At t h i s level the grammar does not include rules of the f o r m "S > NP - VP" or
"NP > NP - PP", but i t does include a l l the rules
which describe the internal structures of BCs at the lowest level of recursion
Every BC has a head and may have one (or more) modifier The head of a BC is the c h a r a c t e r i s t i c word, the word without which a group of words would
f a i l to be an instance of that p a r t i c u l a r BC So the head of a NP is a noun, that of a PP is a pre- position, that of a VP is a verb, e t c ( g ) The head
of a BC carries on a l l the morphological, syntac-
t i c a l and l e x i c a l features of the BC i t s e l f (10) Let us consider the sentence
( I ) "Arturo vide una commedia con Meryl Streep." which may be interpreted both
(1.a) Arthur and Meryl Streep saw a play together and
(1.b) Arthur saw Meryl Streep while she was working
in a play
At t h i s f i r s t level of analysis ( I ) is rewritten
as
PP
v { { / \
(9) The case of PP constitutes a p a r t i a l exception
to t h i s p r i n c i p l e In fact while for syntax is
s u f f i c i e n t to know a l l the relevant informa- tion concerning the preposition, semantics also need to know the information con- cerning the head of the NP which forms the PP (10) This d e f i n i t i o n of head encompasses a l l constructions (endocentric and exocentric);
i t is closer to the t r a d i t i o n a l notion of governing categories than the d e f i n i t i o n given
by Bloomfield [Bloomfield 35] in terms of
d i s t r i b u t i o n See [ M i l l e r 85]
Trang 3The o u t p u t o f t h i s f i r s t s t e p o f s y n t a c t i c a n a l y s i s
i s a s t r u c t u r e t h a t i n c l u d e s t h e s y n t a c t i c a m b i -
g u i t i e s which w i l l be p r o p e r l y t r e a t e d a t t h e
second l e v e l o f a n a l y s i s ( 1 1 )
The second l e v e l o f s y n t a c t i c a n a l y s i s has t h e
g o a l o f s o l v i n g t h e p r o b l e m s a b o u t p r e p o s i t i o n a l
phrase a t t a c h m e n t , noun p h r a s e m o d i f i c a t i o n and
c o n j u n c t i o n and t h a t o f e s t a b l i s h i n g g r a m m a t i c a l
r e l a t i o n s among BCs ( 1 2 ) I n t h e usual s y n t a c t i c
approach t h i s a c t i v i t y , p e r f o r m e d among more
complex c o n s t i t u e n t s , l e a d s t o t h e e x p l o s i o n o f
s t r u c t u r a l a m b i g u i t i e s I n o u r case t h e p r o b l e m o f
h a n d l i n g a m b i g u i t y s t r o n g l y a r i s e s : i n f a c t t h e
s y n t a c t i c a n a l y z e r has been d e s i g n e d i n o r d e r t o
t r e a t a l a r g e v a r i e t y o f r e a l t e x t s which c o n t a i n
words o u t o f t h e i r p r e f e r r e d g r a m m a t i c a l o r d e r o r
which p r e s e n t e l l i p t i c a l c o n s t r u c t i o n s o r , f i n a l l y ,
which p r e s e n t v e r y complex g r a m m a t i c a l c o n s t r u c t s
To r e a c h such an adequacy we r e l a x t h e grammar
c o n s t r a i n t s , b u t t h a t may cause t h e g e n e r a t i o n o f
a r t i f i c i a l s t r u c t u r a l a m b i g u i t i e s ( 1 3 ) In o r d e r t o
s o l v e t h i s p r o b l e m , we see a l l t h e groups o f BCs
h a v i n g t h e same head as b e l o n g i n g t o an e q u i v a l e n c e
c l a s s o f c o n s t i t u e n t s L e t us c o n s i d e r an example
c o n c e r n i n g t h i s i m p o r t a n t p o i n t In I t a l i a n t h e
p h r a s e " I 1 s i n d a c o Rossi d i T o r i n o " ("The m a j o r
Rossi o f T u r i n " ) may i n v o l v e some s t r u c t u r a l ambi-
g u i t y i f i t has t o be p a r s e d w i t h o u t t h e h e l p o f
s e m a n t i c h i n t s I n f a c t , t h i s noun p h r a s e can mean
b o t h t h a t Rossi i s t h e m a j o r o f T u r i n and t h a t
Rossi i s a m a j o r who comes from T u r i n P e r f o r m i n g a
c l a s s i c a l a n a l y s i s t h i s a m b i g u i t y g e n e r a t e s two
d i f f e r e n t s t r u c t u r a l d e s c r i p t i o n s The f i r s t
interpretation can be described as:
NP
(11) At this l e v e l we have not so many ambiguities
because t h e linguis-tic phenomena which
cause them a r e s t i l l n o t f a c e d I n t h i s phase
o f a n a l y s i s l e x i c a l a m b i g u i t y ( i n v o l v i n g
u n c e r t a i n t y a b o u t t h e l e x i c a l c a t e g o r y o f
a g i v e n word) o n l y a r i s e s ; t h i s k i n d o f
a m b i g u i t y i s t r e a t e d by t a k i n g i n t o a c c o u n t
t h e s y n t a g m a t i c r e l a t i o n s h i p s o f t h e words
i n q u e s t i o n ; t h e a n a l y z e r keeps d i f f e r e n t
i n t e r p r e t a t i o n s f o r t h e a m b i g u i t y w h i c h can
n o t be s o l v e d w i t h o u t s e m a n t i c s
(12) Grammatical r e l a t i o n s are p r i m i t i v e notions
such as subject, o b j e c t , complement and so
o n
(13) The constraining power is provided s e t t i n g up
a s t r u c t u r a l homology between s y n t a c t i c and
semantic l e v e l s and performing the formal map-
ping between grammatical r e l a t i o n s and concep-
tual r e l a t i o n s
while the second i n t e r p r e t a t i o n can be described
a s :
NP
PP
f
DI TORINO
In our analysis we handle this problem starting from the consideration that in both the interpreta- tions the NP "Rossi" is the head of the resulting structural unit So the analyzer generates only one representation for the new construction in this way:
SPECIFICATION
Now, let us consider this construction as being part of a sentence:
(2) " I l sindaco Rossi di Torino parte per Roma."
"The major Rossi of Turin is leaving f o r Rome." The a s c r i p t i o n of grammatical r e l a t i o n s among the phrases of t h i s sentence requires the r e c o g n i t i o n
of the NP " l l sindaco Rossi di Torino" as subject
of the sentence and the PP "per Roma" as m o d i f i e r
of the VP The d e t e c t i o n of the subject r e l a t i o n does not necessarily i n v o l v e the problem of s t r u c -
t u r a l ambiguity because t h i s is l i m i t e d at the
r e l a t i o n s between the two NPs and the f i r s t PP So the analyzer gives the f o l l o w i n g d e s c r i p t i o n of the
s e n t e n c e :
SPECIFIC SUBJECT
IL SINDACO C13;?~3 Ol TORINO PARTE PER ROMA
Thanks to this treatment of ambiguity, the syn- tactic structure of this sentence can be described
by o n l y one representation, while a c l a s s i c a l s y n -
t a c t i c analysis would generate at least two repre-
Trang 4sentations Our s i n g l e representation consists of a
graph of BCs connected by grammatical r e l a t i o n s ,
which are established unless s y n t a c t i c knowledge
guarantees t h a t no constituent in the two classes
can be connected by such r e l a t i o n s In t h i s way the
processing is e f f i c i e n t almost as in the case of
complete p a r a l l e l i s m between syntax and semantics
and, in a d d i t i o n , there is complete c o m p a t i b i l i t y
with a p a r a l l e l implementation
Note t h a t none o f the p o s s i b l e i n t e r p r e t a t i o n s
has been l o s t : a l l them are passed t o t h e semantic
i n t e r p r e t e r which o p e r a t e s the r e s o l u t i o n of ambi-
g u i t y t a k i n g i n t o account both the c o n n e c t i o n s b e t -
ween the BCs p o i n t e d o u t by s y n t a c t i c a n a l y s i s and
the semantic p l a u s i b i l i t y o f the proposed connec-
t i o n s
The r e s u l t i n g d i s c r i m i n a t i v e power of syntax is
s t i l l s u f f i c i e n t f o r helping semantics in
e s t a b l i s h i n g the correct semantic r e l a t i o n s among
concepts denoted by words
4 THE SEMANTIC ANALYSIS
Our working hypothesis is t h a t we can represent
the meaning of a t e x t s t a r t i n g from the meanings of
words and from the s y n t a c t i c s t r u c t u r e of the t e x t
We r e p r e s e n t t h e s u r f a c e semantic s t r u c t u r e by
c o n c e p t u a l graphs ( 1 4 ) A c o n c e p t u a l graph i s an
o r i e n t e d b i p a r t i t e graph w i t h two k i n d s o f nodes:
concept nodes ( r e p r e s e n t i n g e n t i t i e s ) and concep-
t u a l r e l a t i o n nodes ( r e p r e s e n t i n g semantic r e l a -
t i o n s among c o n c e p t s ) A Type H i e r a r c h y i s d e f i n e d
o v e r c o n c e p t s
The semantic information is d i s t r i b u t e d on
words by means of canonical graphs, which describe
concepts connoted by the words of the domain in
terms of t h e i r semantic context; they represent the
i m p l i c i t pattern of r e l a t i o n s h i p s necessary f o r a
semantically well-formed t e x t In e a c h canonical
graph we can d i s t i n g u i s h a head (the main concept
node of the canonical graph i t s e l f ) and a semantic
context (see f i g u r e I ) The Type Hierarchy is a
taxonomy of domain concepts used to i n h e r i t seman-
t i c contexts and guide graph j o i n s
The aim o f s u r f a c e semantic a n a l y s i s i s t o
e s t a b l i s h semantic r e l a t i o n s among the head nodes
of c a n o n i c a l graphs connoted by the words of t e x t
F i r s t , the c a n o n i c a l graphs a r e a c t i v a t e d ( c o p i e d
in the w o r k i n g memory); then t h e a c t i v a t e d graphs
are j o i n e d , s u p e r i m p o s i n g c o n t e x t nodes on head
nodes a c c o r d i n g w i t h t h e Type H i e r a r c h y ; so r e l a -
t i o n s are e s t a b l i s h e d among head c o n c e p t s
When e s t a b l i s h i n g a semantic r e l a t i o n , the
mapping with syntax allows the e v a l u a t i o n of i t s
s y n t a c t i c soundness: the s y n t a c t i c a n a l y s i s o u t p u t
(14) The t h e o r y o f Conceptual Graphs i s p r e s e n t e d
by [Sowa, 1984] This f o r m a l i s m i s a g e n e r a l i -
z a t i o n o f v a r i o u s p r e v i o u s approaches t o the
r e p r e s e n t a t i o n of the semantic r e l a t i o n s
h o l d i n g among words such as frames, semantic
networks and c o n c e p t u a l dependency
is checked to see i f a grammatical r e l a t i o n sup- ports the proposed semantic one Otherwise the semantic r e l a t i o n is not established
5 INTEGRATING SYNTAX AND SEMANTICS
During semantic analysis r e l a t i o n s between con- cept nodes are established only i f they are sup-
p o r t e d by t h e r e s u l t o f s y n t a c t i c a n a l y s i s
Given a semantic r e l a t i o n , i t is necessary to see i f there is a corresponding grammatical r e l a -
t i o n The correspondence between grammatical r e l a -
t i o n s and semantic r e l a t i o n s (mapping) is solved through the notion of head which has been i n t r o - duced both in syntax (heads of BCs) and in seman-
t i c s (heads of canonical graphs)
The semantic r e l a t i o n s and the grammatical
r e l a t i o n s must r e l a t e to the same couple of l e x i c a l items; in other words such l e x i c a l items must be both the heads of the BCs (involved by the gram- matical r e l a t i o n ) and the heads of the conceptual graphs (involved by the semantic r e l a t i o n )
A semantic r e l a t i o n SR between two head nodes HNi and HNj, having as heads the words Wi and Wj, can only be established i f :
I) there is a grammatical r e l a t i o n GR between two BCs, BCi and BCj, whose heads are Wi and Wj
r e s p e c t i v e l y 2) semantic r e l a t i o n SR is compatible with the grammatical r e l a t i o n GR and with the set of features Fi and Fj associated to BCi and BCj Conditions are v e r i f i e d through the a p p l i c a t i o n of
a mapping r u l e among a l i m i t e d set Each semantic
r e l a t i o n inside a semantic context of a canonical conceptual graph is augmented with the i n d i c a t i o n
of a mapping r u l e
A mapping r u l e is a l i s t of p l a u s i b l e gram- matical r e l a t i o n s t h a t can correspond to the seman-
t i c r e l a t i o n
In a mapping r u l e each grammatical r e l a t i o n can
be constrained by an a c t i v a t i o n c o n d i t i o n t h a t
r e l a t e s to the morphologic and s y n t a c t i c features
of the involved BC classes
5.1 An example
Let us consider the example of the f i g u r e 2 The j o i n J1 of the head conceptual node HNI with the context node CN2,1 of the head node HN2 causes a conceptual r e l a t i o n AGENT to be established between concept nodes HNI and HN2 Such head concept nodes correspond to words WI ("John") and W2 ("eats") at the l e x i c a l l e v e l
Such conceptual r e l a t i o n has an associated mapping r u l e which requires a grammatical r e l a t i o n
of a c e r t a i n kind (e.g " s u b j e c t " ) Such gram- matical r e l a t i o n must have been established by syn-
t a c t i c analysis between two BCs having WI and W2 as
t h e i r heads As t h a t is the case of f i g u r e 2, the
j o i n J1 can be made
Trang 5D i f f e r e n t l y , j o i n J4 between HN3 and CN2,1 can
n o t be e s t a b l i s h e d as i t would cause an AGENT r e l a -
t i o n between c o n c e p t u a l nodes HN2 ( " e a t " ) and HN3
( " c h i c k e n " ) ; such s e m a n t i c r e l a t i o n i s n o t sup-
p o r t e d by a s u i t a b l e g r a m m a t i c a l r e l a t i o n I n f a c t
there is a grammatical r e l a t i o n between BC2 and
BC3, but i t is not the c o r r e c t one because the
grammatical r e l a t i o n " o b j e c t " can not correspond to
the semantic r e l a t i o n AGENT
To g i v e an i d e a o f t h e mapping r u l e s , t h e
MR-AGENT mapping r u l e i s s k e t c h e d I t i s used t o
map t h e c o n c e p t u a l r e l a t i o n AGENT on t h e gram-
m a t i c a l r e l a t i o n " s u b j e c t " i f t h e a n a l y z e d s e n t e n c e
i s a c t i v e o r on t h e g r a m m a t i c a l r e l a t i o n " a g e n t i v e "
i f t h e s e n t e n c e i s p a s s i v e :
MR-AGENT : s u b j e c t i f BC1 i s ACTIVE and
BCl and BC2 a g r e e
a g e n t i v e i f BC1 i s PASSIVE and
BC2 i s a " b y - p h r a s e "
6 CONCLUSION
The SHEILA system has been p r e s e n t e d as an
a t t e m p t t o s o l v e t h e p r o b l e m o f i n t e g r a t i n g s y n t a x
and s e m a n t i c s The a u t h o r s p r o p o s e t h a t s y n t a c t i c
and s e m a n t i c p r o c e s s e s s h o u l d r e l y on d i s t i n c t
b o d i e s o f knowledge and t h a t t h e i n t e r a c t i o n b e t -
ween s y n t a x and s e m a n t i c s s h o u l d be o b t a i n e d by
e x p l o i t i n g , i n a f o r m a l way, t h e isomorphism b e t -
ween s y n t a c t i c and s e m a n t i c s t r u c t u r e s I n o r d e r t o
a v o i d t h e l a c k o f e f f i c i e n c y c h a r a c t e r i z i n g a
s y n t a x - f i r s t p a r s e r , t h e a u t h o r s have d e s i g n e d a
f l e x i b l e s y n t a x w h i c h , w i t h o u t e x p l o d i n g t h e s t r u c -
t u r a l a m b i g u i t i e s , s u p p l i e s s e m a n t i c i n t e r p r e t e r
w i t h knowledge a b o u t s y n t a c t i c c o n n e c t i o n s between
t h e words o c c u r r i n g i n t h e t e x t The isomorphism
between s y n t a x and s e m a n t i c s i s a c c o u n t e d i n t o a
l i m i t e d s e t o f f o r m a l mapping r u l e s and c o n d i t i o n s
P r e p o s i t i o n a l p h r a s e a t t a c h m e n t , a p p o s i t i o n , d e t e r -
m i n a t i o n o f c o n j u n c t i o n ' s scope and m o d i f i c a t i o n o f
a NP t h r o u g h o t h e r NPs a r e d e a l t i n a s a t i s f a c t o r y
way b o t h from a s y n t a c t i c a l and from a s e m a n t i c a l
p o i n t o f v i e w O t h e r complex l i n g u i s t i c phenomena
(as a n a p h o r a , q u a n t i f i c a t i o n and e l l i p s i s ) r e q u i r e s
a more e x t e n s i v e use o f h e u r i s t i c s The f u t u r e work
w i l l concentrate on these s p e c i f i c aspects in order
to check the adequacy of the hypothesis of i s o - morphism between s y n t a c t i c and semantic s t r u c t u r e s
to l a r g e r fragments of the I t a l i a n language
REFERENCES
[ B a y e r 85] B a y e r , S , Joseph, L , K a ] i s h , C , Grammatical R e l a t i o n s as The B a s i s f o r NL P a r s i n g and T e x t U n d e r s t a n d i n g P r o c 9 t h I J C A I , Los
A n g e l e s , 1985, pp ?98-?90
[Birnbaum 81] Birnbaum, L and S e l f r i d g e , M.,
C o n c e p t u a l A n a l y s i s f o r Language i n Schank, R.C and R i e s b e c k , C K , ( e d s ) , I n s i d e Computers
U n d e r s t a n d i n g Lawrence Erlbaum A s s , 1981
[ B l o o m f i e l d 35] B l o o m f i e l d , L e o n a r d , Language,
A l l e n & Unwin, London 1935
[ F r e d e r k i n g 85] F r e d e r k i n g , R E , S y n t a x and Semantics i n NL P a r s e r s T e c h n i c a l R e p o r t 133,
C a r n e g i e - M e l l o n , D e p t o f Computer S c i e n c e , May
1985
[ H i r s t 84] H i r s t , G J , S e m a n t i c I n t e r p r e t a t i o n
A g a i n s t A m b i g u i t y Brown U n i v e r s i t y , P h D , 1984 [Lesmo 85] Lesmo, L and T o r a s s o , P W e i g h t e d ,
I n t e r a c t i o n o f S y n t a x and Semantics i n NL A n a l y s i s Pro¢ 9 t h I J C A I , Los A n g e l e s , 1985, ?72-??8
[ L y t i n e n 85] L y t i n e n , S L , I n t e g r a t i n g S y n t a x and S e m a n t i c s , P r o c T h e o r e t i c a l and
M e t h o d o l o g i c a l I s s u e s i n MT f o r NLs, H a m i l t o n ,
1985, 1 6 7 - 1 7 8
[ M i l l e r 85] M i l l e r , J , Semantics and Syntax,
Cambridge Univ Press, Cambridge (U.K.), 1985 [Stock 86] Stock, 0 , Dynamic U n i f i c a t i o n in
L e x i c a l l y Based Parsing., Proc ?th ECAI, Brighton,
1986, 212-221
[Sowa 84] Sowa, J F , C o n c e p t u a l S t r u c t u r e s Addison Wesley, 1984
[Woods ?2] Woods, W.A., An E x p e r i m e n t a l P a r s i n g System f o r T r a n s i t i o n Network Grammars T e c h n i c a l
R e p o r t 2362, B o l t Beranek and Newman I n c , 1972
Trang 6CONCEPTS CANONICAL GRAPHS TYPE HIERARCHY
( ~ " ~' SEN CONTEXT
~.Co, J)'-d POOO 1 " ~ , ~ I I ~.s' "
SEN CONTEXT
SURFACE SEMANTIC LEVEL
LEXICAL
JOHN EATS CHICKEN FORK LEVEL
W_ I
J~ I ~ I ~ I ~ SYNTACTIC I~ ~ I ~ ~ T ] ~ LEVEL
COMPLEMENT
I ) JOHN I ) EATS I ) ACHICKEN I ) WITH A FORK
2) A CHICKEN WITH A FORK
F i g 1 - T h e c a n o n i c a ] graph o f " e a t " and t h a t o f
" f o r k "
F i g 2 - Mapping aspects f o r t h e sentence "John e a t s
a c h i c k e n w i t h t h e f o r k " The s y n t a c t i c l e -
v e l r e p r e s e n t s t h e graph o f BCs t h a t con-
s t i t u t e s t h e two s y n t a c t i c s t r u c t u r e s o f
t h e s e n t e n c e At the semantic l e v e l d o t t e d arrows ( ~ ) stand f o r a j o i n t h a t i s
s u p p o r t e d by s y n t a x The d o u b l e arrows ( C > ) i n s t e a d r e p r e s e n t s a j o i n t h a t i s
n o t supposed by s y n t a x In f a c t a mapping
r u l e r e q u i r e s t h a t t h e semantic r e l a t i o n
" a g e n t " must be s u p p o r t e d by t h e g r a m m a t i -
c a l r e l a t i o n " s u b j e c t " ( i n an a c t i v e sen- tence) and not by the " o b j e c t " r e l a t i o n ,