<?xml version="1.0" standalone="yes"?>
<Paper uid="E87-1044">
  <Title>T INTEGRATING SEMANTICS kNO FLEXIBLE SYNTAX BY EXPLOITING ISONORPHISM BETWEEN GRAIelATICAL AND SEMANTICAL RELATIONS</Title>
  <Section position="2" start_page="0" end_page="0" type="metho">
    <SectionTitle>
* ABSTRACT
</SectionTitle>
    <Paragraph position="0"> This work concerns integration between syntax and semantics. Syntactic and semantic activities rely on separate bodies of knowledges. Integration is obtained by exploiting the isomorphism between grammatical relations (among immediate constitu~ errs) and conceptual relations, thanks to a limited set of formal mapping rules. Syntactic analysis does not construct all the explicit parse trees but just a graph that represents all the plausible grammatical relations among immediate constituents.</Paragraph>
    <Paragraph position="1"> Such graph gives the semantic interpreter, based on Conceptual Graphs formalism, the discriminative power required to establish conceptual relations.</Paragraph>
  </Section>
  <Section position="3" start_page="0" end_page="0" type="metho">
    <SectionTitle>
I. INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> In the field of automatic natural language understanding, the problem of connecting syntax and semantics has been faced in three different ways.</Paragraph>
    <Paragraph position="1"> Some authors are persuaded that understanding natural language requires no use of syntactic knowledge. They claim that semantic representation can be built directly from the surface string, without the help of almost any syntactic source (1).</Paragraph>
    <Paragraph position="2"> Other authors proposed highly syntactic systems, starting from the idea that the representation of the syntactic structure is preliminary to the understanding process (2).</Paragraph>
    <Paragraph position="3"> While the work of this second group of researchers was concerned mainly with the understanding of individual sentences, the work of the partisans of semantics was about the understanding of whole texts.</Paragraph>
    <Paragraph position="4"> This shifting of attention substained the idea that syntax and semantics should be used in an integrated way. Most researchers have thought that semantics and syntax should be integrated with respect to both the representation and the processing (3); others have claimed that it is more efficient to build a full-blooded syntactic representation during the parsing process (4).</Paragraph>
    <Paragraph position="5">  (1) See the system IPP \[Schank 80\].</Paragraph>
    <Paragraph position="6"> (2) The LUNAR system \[Woods ?2\] is a classical example.</Paragraph>
    <Paragraph position="7"> (3) An example is the Conceptual Analyzer \[Birnbaum 81\].</Paragraph>
    <Paragraph position="8"> (4) See MOPTRANS \[Lytinen 85\],  Our approach shares some communalities with the last position. We reckon that semantic and syntactic processes should rely on separate knowledge bodies. Uur effort is mainly focused on the realization of the integration by exploiting the isomorphism between syntactic structures and semantic representations, rather than by making syntactic and semantic processes interact, as it happens in previous integrated parsers (5). The idea of isomorphism is not carried out through one-to-one correspondence between syntactic rules and semantic ones - as in Montague-inspired parsers (6), but by mapping in a formal way grammatical and conceptual relations. The use of grammatical relations as intermediate level between syntax and semantics was also adopted in the KING KONG parser (7), but this system is still more near to the position which wants the representation of syntax and semantics as well as their processes to interact, while our choice is to maintain separate these different sources of knowledge.</Paragraph>
    <Paragraph position="9"> The subsequent paragraphs describe how this hypothesis works in SHEZLA (Syntax Helping Expectations In Language Analysis), a prototype developed at CSELT laboratories (Turin, Italy). The aim of SHEILA is to analyze and to extract relevant information from news (coming from the Italian news agency &amp;quot;ANSA&amp;quot;). The system is initially being applied to texts describing variations in the topmanagement of commercial societies; it has been fully implemented on a Symbolics Lisp machine.</Paragraph>
    <Paragraph position="10"> SHEILA takes advantage both from the use of expectations and from the combination of the results of a non-conventional syntactic analysis with the activity of a surface semantic analysis, based on the formalism of conceptual graphs (8). In this paper we describe just the principles which guide the integration between syntax and semantics.</Paragraph>
    <Paragraph position="11"> SHEILA correctly analyzes a set of thirty news, generating for each of them a set of records for a relational data base.</Paragraph>
  </Section>
  <Section position="4" start_page="0" end_page="278" type="metho">
    <SectionTitle>
2. THE PROBLEM AND OUR PROPOSAL
</SectionTitle>
    <Paragraph position="0"> In text understanding systems syntax and semantics have almost always been dealt with integration of their processing. Usually this kind of  (5) See PSLI3 \[Frederking 85\], FZDO \[Lesmo 85\] and WEDNESDAY-2 \[Stock 86\].</Paragraph>
    <Paragraph position="1"> (6) See ABSITY \[Hirst 84\].</Paragraph>
    <Paragraph position="2"> (7) See \[BAYER 65\] (8) See \[Sowa 84\] and also the fourth paragraph below. null  systems are semantic driven and they do only local syntactic checks during analysis. Doing local syntactic checks only involves little amount of syntactic knowledge and that is misleading in solving problems as anaphoric reference, prepositional attachment, conjunction and so on.</Paragraph>
    <Paragraph position="3"> In a different approach the integration has been realized during the syntactic structure representation construction: the syntactic parser makes use of semantic information to handle structural ambiguities.</Paragraph>
    <Paragraph position="4"> The questioning done by the syntactic analyzer to the semantic component a~ms to cut down the number of parse trees, but very many rules are required for this questioning, which has always been the most domain-dependent part of natural language understanding systems.</Paragraph>
    <Paragraph position="5"> In designing SHEILA we chose another way of integrating syntax with semantics. The basic schema may look rather classic: the system produces a syntactic analysis of the text, driven on the basis of purely syntactic knowledge. The semantic analyzer checks the syntactic output to see if the semantic relations among words are supported by it.</Paragraph>
    <Paragraph position="6"> But a classical syntax-first analysis is highly inefficient. It cannot solve structural 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 artificial syntactic ambiguities. So there are two problems: reducing the explosion of ambiguities and determining how semantic patterns for each word interact with syntax.</Paragraph>
    <Paragraph position="7"> Our proposal faces these problems through the original combination of two key ideas, i.e.: I) a flexible syntactic analysis, which is performed by constructing not all the explicit parse trees but just a graph, representing all the plausible grammatical relations among immediate constituents; 2) a formal way of interaction between syntax and semantics exploiting the isomorphism between syntactic structures (grammatical relations among immediate constituents) and semantic ones (conceptual relations).</Paragraph>
    <Paragraph position="8"> Such flexible syntactic analysis gains a discriminative power (sufficient for aiding semantics in solving ambiguities) and avoids the explosion in the parse trees number. Furthermore, the mapping between grammatical and conceptual relations can be defined through a limited set of formal rules.</Paragraph>
  </Section>
  <Section position="5" start_page="278" end_page="278" type="metho">
    <SectionTitle>
3. THE SYNTACTIC ANALYSIS
</SectionTitle>
    <Paragraph position="0"> Our system has the goal of generating a semantic structure that has to be consistent with the syntactic form used to convey it in the text. The aim of syntactic analysis is to support semantics.</Paragraph>
    <Paragraph position="1"> A first activity performed by the syntactic @nalyzer is the recognition of constituents of the phrase structure of text. This is done by applying a set of rewriting phrase structure rules for Italian language. These rules utilize the output of a previous morphological analysis that assigns to words morphological and lexical features (gender, number, lexical category and so on).</Paragraph>
    <Paragraph position="2"> In this 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 this level the grammar does not include rules of the form &amp;quot;S --&gt; NP - VP&amp;quot; or &amp;quot;NP --&gt; NP - PP&amp;quot;, but it does include all the rules which describe the internal structures of BCs at the lowest level of recursion.</Paragraph>
    <Paragraph position="3"> Every BC has a head and may have one (or more) modifier. The head of a BC is the characteristic word, the word without which a group of words would fail to be an instance of that particular BC. So the head of a NP is a noun, that of a PP is a preposition, that of a VP is a verb, etc.(g). The head of a BC carries on all the morphological, syntactical and lexical features of the BC itself (10). Let us consider the sentence  (I) &amp;quot;Arturo vide una commedia con Meryl Streep.&amp;quot;. 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.</Paragraph>
    <Paragraph position="4"> At this first level of analysis (I) is rewritten as</Paragraph>
    <Paragraph position="6"/>
  </Section>
  <Section position="6" start_page="278" end_page="280" type="metho">
    <SectionTitle>
ARTURO VIDE UNA COMMEDIA CON MERYL STREEP
</SectionTitle>
    <Paragraph position="0"> (9) The case of PP constitutes a partial exception to this principle. In fact while for syntax is sufficient to know all the relevant information concerning the preposition, semantics also need to know the information concerning the head of the NP which forms the PP.</Paragraph>
    <Paragraph position="1"> (10) This definition of head encompasses all constructions (endocentric and exocentric);  it is closer to the traditional notion of governing categories than the definition given by Bloomfield \[Bloomfield 35\] in terms of distribution. See \[Miller 85\].</Paragraph>
    <Paragraph position="2">  The output of this first step of syntactic analysis is a structure that includes the syntactic ambiguities which will be properly treated at the second level of analysis (11).</Paragraph>
    <Paragraph position="3"> The second level of syntactic analysis has the goal of solving the problems about prepositional phrase attachment, noun phrase modification and conjunction and that of establishing grammatical relations among BCs (12). In the usual syntactic approach this activity, performed among more complex constituents, leads to the explosion of structural ambiguities. In our case the problem of handling ambiguity strongly arises: in fact the syntactic analyzer has been designed in order to treat a large variety of real texts which contain words out of their preferred grammatical order or which present elliptical constructions or, finally, which present very complex grammatical constructs. To reach such an adequacy we relax the grammar constraints, but that may cause the generation of artificial structural ambiguities (13). In order to solve this problem, we see all the groups of BCs having the same head as belonging to an equivalence class of constituents. Let us consider an example concerning this important point. In Italian the phrase &amp;quot;I1 sindaco Rossi di Torino&amp;quot; (&amp;quot;The major Rossi of Turin&amp;quot;) may involve some structural ambiguity if it has to be parsed without the help of semantic hints. In fact, this noun phrase can mean both that Rossi is the major of Turin and that Rossi is a major who comes from Turin. Performing a classical analysis this ambiguity generates two different structural descriptions. The first interpretation can be described as:</Paragraph>
    <Paragraph position="5"> (11) At this level we have not so many ambiguities because the linguis-tic phenomena which cause them are still not faced. In this phase of analysis lexical ambiguity (involving uncertainty about the lexical category of a given word) only arises; this kind of ambiguity is treated by taking into account the syntagmatic relationships of the words in question; the analyzer keeps different interpretations for the ambiguity which can not be solved without semantics.</Paragraph>
    <Paragraph position="6">  (12) Grammatical relations are primitive notions such as subject, object, complement and so on.</Paragraph>
    <Paragraph position="7"> (13) The constraining power is provided setting up  a structural homology between syntactic and semantic levels and performing the formal mapping between grammatical relations and conceptual relations.</Paragraph>
    <Paragraph position="8"> while the second interpretation can be described as:</Paragraph>
    <Paragraph position="10"> In our analysis we handle this problem starting from the consideration that in both the interpretations the NP &amp;quot;Rossi&amp;quot; is the head of the resulting structural unit. So the analyzer generates only one representation for the new construction in this way:</Paragraph>
    <Paragraph position="12"> Now, let us consider this construction as being part of a sentence: (2) &amp;quot;Il sindaco Rossi di Torino parte per Roma.&amp;quot; &amp;quot;The major Rossi of Turin is leaving for Rome.&amp;quot; The ascription of grammatical relations among the phrases of this sentence requires the recognition of the NP &amp;quot;ll sindaco Rossi di Torino&amp;quot; as subject of the sentence and the PP &amp;quot;per Roma&amp;quot; as modifier of the VP. The detection of the subject relation does not necessarily involve the problem of structural ambiguity because this is limited at the relations between the two NPs and the first PP. So the analyzer gives the following description of the sentence:</Paragraph>
    <Paragraph position="14"> Thanks to this treatment of ambiguity, the syntactic structure of this sentence can be described by only one representation, while a classical syntactic analysis would generate at least two repre- null sentations. Our single representation consists of a graph of BCs connected by grammatical relations, which are established unless syntactic knowledge guarantees that no constituent in the two classes can be connected by such relations. In this way the processing is efficient almost as in the case of complete parallelism between syntax and semantics and, in addition, there is complete compatibility with a parallel implementation.</Paragraph>
    <Paragraph position="15"> Note that none of the possible interpretations has been lost: all them are passed to the semantic interpreter which operates the resolution of ambiguity taking into account both the connections between the BCs pointed out by syntactic analysis and the semantic plausibility of the proposed connections. null The resulting discriminative power of syntax is still sufficient for helping semantics in establishing the correct semantic relations among concepts denoted by words.</Paragraph>
  </Section>
  <Section position="7" start_page="280" end_page="280" type="metho">
    <SectionTitle>
4. THE SEMANTIC ANALYSIS
</SectionTitle>
    <Paragraph position="0"> Our working hypothesis is that we can represent the meaning of a text starting from the meanings of words and from the syntactic structure of the text.</Paragraph>
    <Paragraph position="1"> We represent the surface semantic structure by conceptual graphs (14). A conceptual graph is an oriented bipartite graph with two kinds of nodes: concept nodes (representing entities) and conceptual relation nodes (representing semantic relations among concepts). A Type Hierarchy is defined over concepts.</Paragraph>
    <Paragraph position="2"> The semantic information is distributed on words by means of canonical graphs, which describe concepts connoted by the words of the domain in terms of their semantic context; they represent the implicit pattern of relationships necessary for a semantically well-formed text. In each canonical graph we can distinguish a head (the main concept node of the canonical graph itself) and a semantic context (see figure I). The Type Hierarchy is a taxonomy of domain concepts used to inherit semantic contexts and guide graph joins.</Paragraph>
    <Paragraph position="3"> The aim of surface semantic analysis is to establish semantic relations among the head nodes of canonical graphs connoted by the words of text.</Paragraph>
    <Paragraph position="4"> First, the canonical graphs are activated (copied in the working memory); then the activated graphs are joined, superimposing context nodes on head nodes according with the Type Hierarchy; so relations are established among head concepts.</Paragraph>
    <Paragraph position="5"> When establishing a semantic relation, the mapping with syntax allows the evaluation of its syntactic soundness: the syntactic analysis output (14) The theory of Conceptual Graphs is presented by \[Sowa, 1984\]. This formalism is a generalization of various previous approaches to the representation of the semantic relations holding among words such as frames, semantic networks and conceptual dependency.</Paragraph>
    <Paragraph position="6"> is checked to see if a grammatical relation supports the proposed semantic one. Otherwise the semantic relation is not established.</Paragraph>
  </Section>
  <Section position="8" start_page="280" end_page="281" type="metho">
    <SectionTitle>
5. INTEGRATING SYNTAX AND SEMANTICS
</SectionTitle>
    <Paragraph position="0"> During semantic analysis relations between concept nodes are established only if they are supported by the result of syntactic analysis.</Paragraph>
    <Paragraph position="1"> Given a semantic relation, it is necessary to see if there is a corresponding grammatical relation. The correspondence between grammatical relations and semantic relations (mapping) is solved through the notion of head which has been introduced both in syntax (heads of BCs) and in semantics (heads of canonical graphs).</Paragraph>
    <Paragraph position="2"> The semantic relations and the grammatical relations must relate to the same couple of lexical items; in other words such lexical items must be both the heads of the BCs (involved by the grammatical relation) and the heads of the conceptual graphs (involved by the semantic relation).</Paragraph>
    <Paragraph position="3"> A semantic relation SR between two head nodes HNi and HNj, having as heads the words Wi and Wj, can only be established if: I) there is a grammatical relation GR between two BCs, BCi and BCj, whose heads are Wi and Wj respectively.</Paragraph>
    <Paragraph position="4"> 2) semantic relation SR is compatible with the grammatical relation GR and with the set of features Fi and Fj associated to BCi and BCj.</Paragraph>
    <Paragraph position="5"> Conditions are verified through the application of a mapping rule among a limited set. Each semantic relation inside a semantic context of a canonical conceptual graph is augmented with the indication of a mapping rule.</Paragraph>
    <Paragraph position="6"> A mapping rule is a list of plausible grammatical relations that can correspond to the semantic relation.</Paragraph>
    <Paragraph position="7"> In a mapping rule each grammatical relation can be constrained by an activation condition that relates to the morphologic and syntactic features of the involved BC classes.</Paragraph>
    <Section position="1" start_page="280" end_page="281" type="sub_section">
      <SectionTitle>
5.1 An example
</SectionTitle>
      <Paragraph position="0"> Let us consider the example of the figure 2.</Paragraph>
      <Paragraph position="1"> The join J1 of the head conceptual node HNI with the context node CN2,1 of the head node HN2 causes a conceptual relation AGENT to be established between concept nodes HNI and HN2. Such head concept nodes correspond to words WI (&amp;quot;John&amp;quot;) and W2 (&amp;quot;eats&amp;quot;) at the lexical level.</Paragraph>
      <Paragraph position="2"> Such conceptual relation has an associated mapping rule which requires a grammatical relation of a certain kind (e.g. &amp;quot;subject&amp;quot;). Such grammatical relation must have been established by syntactic analysis between two BCs having WI and W2 as their heads. As that is the case of figure 2, the join J1 can be made.</Paragraph>
      <Paragraph position="3">  Differently, join J4 between HN3 and CN2,1 can not be established as it would cause an AGENT relation between conceptual nodes HN2 (&amp;quot;eat&amp;quot;) and HN3 (&amp;quot;chicken&amp;quot;); such semantic relation is not supported by a suitable grammatical relation. In fact there is a grammatical relation between BC2 and BC3, but it is not the correct one because the grammatical relation &amp;quot;object&amp;quot; can not correspond to the semantic relation AGENT.</Paragraph>
      <Paragraph position="4"> To give an idea of the mapping rules, the MR-AGENT mapping rule is sketched. It is used to map the conceptual relation AGENT on the grammatical relation &amp;quot;subject&amp;quot; if the analyzed sentence is active or on the grammatical relation &amp;quot;agentive&amp;quot; if the sentence is passive: MR-AGENT : subject if BC1 is ACTIVE and BCl and BC2 agree.</Paragraph>
      <Paragraph position="5"> agentive if BC1 is PASSIVE and BC2 is a &amp;quot;by-phrase&amp;quot;</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>