<?xml version="1.0" standalone="yes"?>
<Paper uid="C96-2137">
  <Title>Anaphora Resolution of Japanese Zero Pronouns with Deictic Reference</Title>
  <Section position="4" start_page="0" end_page="812" type="metho">
    <SectionTitle>
2 Appearance of Zero Pronouns in
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="0" end_page="812" type="sub_section">
      <SectionTitle>
Japanese Texts
</SectionTitle>
      <Paragraph position="0"> In order to understand the distribution of zero pronouns with antecedents that do not appeal&amp;quot; in the text, in this section, we examine which zero pronouns nmst be resolved and where their antecedents appear, using a test set designed to evaluate the performance of Japanese-to-English ma- null chine translation systems (lkehara et al., 1994). 'l'he results of the examination of zero pronouns and their referential elements in the flmctional test sentence set (3718 sentences) are shown in Table l. There were ~, total of 512 zero pronomls in 463 sentences. The location of referential elements can be divided into 2 kinds: those expressed in the same sentence, and those, not expressed in the same sentence. 'rim latter were fltrther classitied into 6 kinds.</Paragraph>
      <Paragraph position="1">  According to this study of the fimctional test sentence set, in 373 out of 512 instances (73%) the antecedent was not expressed in the sentence. Zero pronouns conhl be left unexpressed by converting the translation to the passive voice in 173 instances (34%). 'Fhe other zero pron(mns, 200 instances (39%), referred to antecedents that (lid not appear in the sentence. \[n 69 ()tit, of the 200 instances (13%) zero pronouns were th.e subject of the sentence and referred to the writer or speaker I or a group we. Further examination revealed that only in these 69 instances did the verb that governed them express some modality such as shilai '- want to -' or -shiyou 'Let us -' or the verbs were omou 'tMnk' and other such words indicating ~TI\]\[INKING ACTION'. Furthermore, zero pronouns that were the subjects and that referred to the reader or hearer you, ainounted to 28 out of the 200 instances (5%). In these 28 instances, ~he verbs that governed these zero pronouns expressed the modalities of- subekida 'should' or -sitehanaranai 'nmst n(it'. Similarly, modalities and verb types can be used to identify it or the 'unknown human'. This tyl)e of zero t)ron(mn can be resolved by deducing their referents using modality or categor:ized verl)al semantic attrilmtes. null</Paragraph>
    </Section>
  </Section>
  <Section position="5" start_page="812" end_page="815" type="metho">
    <SectionTitle>
3 Deictic Resolution of Japanese
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="812" end_page="812" type="sub_section">
      <SectionTitle>
Zero Pronouns
</SectionTitle>
      <Paragraph position="0"> lhused on the results shown in section 2, we propose a method to resolve Japanese zero pronouns whose antecedents do not appear in the texts.</Paragraph>
    </Section>
    <Section position="2" start_page="812" end_page="812" type="sub_section">
      <SectionTitle>
3.1 Dei('tle Resolution using Semantic
Constraints on Cases
</SectionTitle>
      <Paragraph position="0"> ~lb resolve Japanese zero pronouns whose antecedents do not appear within the texts, it is possible to use the semantic constraints on verbs' case</Paragraph>
    </Section>
    <Section position="3" start_page="812" end_page="812" type="sub_section">
      <SectionTitle>
referential elements
</SectionTitle>
      <Paragraph position="0"> ~~ Loc. of 'referential elements' \[ -,t- II p~w I l~-yo.--7\]- ~,, in. I t~l elements to deduce like.ly referents. The semantic information used to estimate supplementing elements is similar to the constraints on cases used tbr selecting the transfer patterns in a machine translation systeln. Figure \[ shows an examph: of a transtb.r pattern in a Japanese-to-English tnachine translation syst, em for the ,lapanese~ verb ikimasu 'gC. Figure 1 shows how, if the ,lapane.se verb is ikimasu 'go' and the noun phr~use with a ga pc.riMe, which shows a subject, has the senmntic attribute SUBJECT, VEIIICLES OR, ANIMALS~ then the verb should be translated as 'go'. In this pattern, if Lit(; suhject N I beeonles a zero l)ronoun, the system tries to estimate the referent using semantic constraints. But, in this case., it is impossible to estimate the referent as one type, be.cause there are three kinds of semantic constraints. In the transfer pattern, the semantic constraints are left unfulfilled if they are not used in selecting the appropriate translation. So, this method frequently l)oses di\[\[iculties in pinpointing elements to be estimated.</Paragraph>
      <Paragraph position="1"> According to die results that were examined in section 2, this type of zero pronoun can be resolved by deducing I;heir reDrents not only /tsing semantic constraints to the cases but also using modality or categorized verbal semantic ato tributes. For example, in this case, it is efl'cetive to determine the referents corresponding to 'I' itsing the verbal semantic attributes of the pattern, N \['S PIIYSI(JAL TRANSI,'ER and the polite exl)ression -maser.</Paragraph>
      <Paragraph position="3"/>
    </Section>
    <Section position="4" start_page="812" end_page="813" type="sub_section">
      <SectionTitle>
3.2 Deietic Re, solution using Semantic
and Pragmatic Constraints
</SectionTitle>
      <Paragraph position="0"> According to the analysis of the results shown in section 2, we found that modal expressions and verbal semantic a.ttribntes are usefltl in (leretraining the dei('tic referents of Japanese zero pronouns. Also, we can estimate the types of conjun&lt;:tions that are effe(:tive in determining the tel&lt;  erents in a complex sentence. In this section, we examine three kinds of semantic and pragmatic constraints, modal expressions, verbal semantic attributes arid conjunctions.</Paragraph>
      <Paragraph position="1">  Modal expressions in Japanese are expected Co be the most powerful constraints for estimating deictie reference. For example, in the case of zero prononns in ga-cases 'subject', the referent becomes the writer or speaker, I or a group, we if the sentence has the modal expressions, -sitai '4) want to -' HOPF, or -silchosii 'C/ want C/ to-' CAUSATIVE HOI'E; the referent becomes the reader or hearer, you if the sentence has the modal expressions, sileha-ikenai '4) must not -' PR()IIIBIT or -subekida 'C/ should -' OBLmA'rmN. If there are no refereat candidates found within the surrounding text, the referents can be determined using the previous constraints based on modal expressions.</Paragraph>
    </Section>
    <Section position="5" start_page="813" end_page="814" type="sub_section">
      <SectionTitle>
Smnantie Attributes
</SectionTitle>
      <Paragraph position="0"> Constraints b~sed on verbal semantic attributes can be divided into tile \[bllowing two types:  (1) Constraints based on the types of verbs 'Give and take' expressions such as the verbs moran 'get' and yaru 'give' and transfer expres null sions such as the verbs iku 'go' and h'ur'u 'come' can deternfine the referents of zero pronouns without modal expressions. For example, if the ga-case (subject) of the sentence whose verb is ,nora'tt 'get' becolnes a zero pronoun, the referent becomes 1. In the case of verb kuru 'come', the referent becomes an element other than i, tbr example you. These kinds of verbs inrplicitly indicate tile relationship between the writer/speaker and the referent of the ga-case (for example, the empathy (Kuno, 1978) or the side of the territory of infbrmation (Kmnio, 1985)). Based on these properties, tire deictic referents of Japanese zero pronouns (:an be estimated.</Paragraph>
      <Paragraph position="1"> (2) Constraints based on the types of verbs and modal expressious I,;ven if the referents of zero pronouns (:an not be determined using modal expressions or tile types of verbs, the referents can sometimes be determined using a combination of modal expressions and the types of verbs. For example, it1 the following Japanese expression, the ga-case becomes a zero pronoun.</Paragraph>
      <Paragraph position="3"> I read a book.</Paragraph>
      <Paragraph position="4"> In this sentence the experience of the writer/speaker, 1 is suitable for the reference of the zero pronoun. As shown in this sentence, if the ga-c~use in an expression with a verb whose semantic attribute is ACTION and modal expre.ssion is -ta PAST, becomes a zero pronoun, it will be translated by a truman translator ~s 1. In a similar way, if the ga-case in an expression with a verb whose semantic attribute is ACT1ON and modal expression is -darou 'will' FSTIMA'FION, becomes a zero l)ronottn, the referent is you. Such constraints using both verbal semantic attributes and modal expressions can be used to determine the deictic reference of 3 apanese zero pronouns.</Paragraph>
      <Paragraph position="5"> To write constraints based on types of verbs eflbx:tively, we used the 97 verbal semantic attributes (VSA) proposed by Nakaiwa (1994).</Paragraph>
      <Paragraph position="6">  Sometimes the deictic referents of Japanese zero pronouns can be determined depending on the types of conjunctions. The constraints based on the Japanese conjunctions can be divided into the following two types.</Paragraph>
      <Paragraph position="7"> (1) The eonstralnts on &lt;:as(; sharing depending on the types of conjunctions Minami (1974) and Takubo (1987) proposed that (lilt'event Japanese conjunctions cover or share different cases, l,'or example Minanfi divided Japanese conjnnctions into three kinds, A, B and C. A complex sentence which includes A type Japanese conjunctions, such as tsu-Lsu 'while' and nagara 'while', shares one ha-case ('l'opie) and one ga-ea,se (Subject). In the case of B type Japanese conjullctions, stlch a8 lzodc 'because' or tara 'iF, one ha.-c~e is shared but not the ga-case. In the case of C type J a I)an esc conj u actions, such as kcredo 'but' or kedo 'but', neither the ha-case nor the ga-case are necessarily shared. According to this classification, if two ga-cases in a complex sin&gt; tence joined by an A type Japanese conjunction were to become zero pronouns and the referent of one of the two zero pronouns wins determined by the constraints proposed previously, then the referent of the other zero pronoun is the same referent. These characteristics of Japanese c(mjunctions can be used to determine the refit'rents of zero pronouns.</Paragraph>
      <Paragraph position="8"> (2) Constraints based on conjunctions, inodal expressions and verbal semantic attributes null Sometimes co-occurrence of conjunctions, verbal semantic attributes and moda.l expressions in a complex sentence determines the meaning of the sentence, and sometimes they determine the deictic referen(;e of zero pronouns in the sentence. For example, in the following Japanese expression, th(~ subject of the verb ika-nai 'go-not' becomes a zero pronoun but the referent can be determin(;d as the writer or speaker, you.</Paragraph>
      <Paragraph position="9"> (2) tokoya-ni ika-nai to, C/-SUBJ b~'ber-IND-O13 J go-not if If you don't go to the tmrber, kami-ga boubou-ni-naru hair begin to look unt2dy your hair will begin to look untidy.</Paragraph>
      <Paragraph position="10">  This sentence has the meaning that the writer or speaker advises that if you do not do something, a situation will arise. The meaning type of a complex sentence can be determined using the rules that the eOBjlIIICtiO\[I is tO 'if' and in the sub clause ga-case beeoInes a zero pronoun and tile meaning; o\[' the verb is ACTION with negation and in the main clause the meaning of the verb is A'P'PlLIBU'FE with tnoda\] expression ni-uaru 'become' AT'I'ItlB U'\['\],', '\['ltA N S FE \[1..</Paragraph>
      <Paragraph position="11"> '\['he meaning type of a complex sentence can he determined using the tblowing ruh'.s: when the con.imletion is 'if' ~md the sill) clause ga-c;~se be(;on~es a zero \]\]FOlIO/Ill, and the meanil~g of the verb is aCTION with negation, and in the main cla.use the meaning of the verb is ATTRIBUT\],; with modal expression, then ni-naru q)eeome' is an example of A'\[&amp;quot;PI{IBUTE TRANSI,'I~3R. Using these kinds of rules, the meaning types o1&amp;quot; complex sentences can I)e determined, and the reference of zero pr(m(mns c.an be deterufine.d.</Paragraph>
    </Section>
    <Section position="6" start_page="814" end_page="814" type="sub_section">
      <SectionTitle>
3.3 Algorithm
</SectionTitle>
      <Paragraph position="0"> In this sul)section, we t)ropose an algorithm for the deictic resolution of Japanese zero pronouns using the constraints proposed in this section.</Paragraph>
      <Paragraph position="1"> This algorithm was implemented in a Japanesetod,3nglish machine translation system, so the only zero pron(mns that must be resolved are those tlu~t become mandatory elernents in li\]nglish. To realize the previously proposed conditions in an algorithm, we must consider eases when these tmtece(lelll, s exist in the same selitellCe ;ts well as when these antecedents exist in another sent, ences in the text, and we must design the algorithm to increase the eve.rail accuracy of the resolution of zero pronouns.</Paragraph>
      <Paragraph position="2"> Anaphora resolution of zero pronouns is conducted as follows. In each step in the algorithm, when the referential element within or without the text is determined, the system checks not only the conditions that are written in the following algorithm, I)ut also the semantic conditions that verbs impose on zero pronouns in tile case elements in each pattern of the Japanese-to-English transfer dictionaries.</Paragraph>
      <Paragraph position="3"> l) \])etection of zero pronouns.</Paragraph>
      <Paragraph position="4"> If they exist, proceed to step 2.</Paragraph>
      <Paragraph position="5"> 2) \[CXttlttilte whether there ~r(.&amp;quot; ~ntecedents within the same sentelwes. (\]&amp;quot;or ex;mtple, a.n~tl)\[tora. rcsohttion is performed using Nakaiw~t's method (N,tkaiwtt and Ikehara, 1995)), If their antecedents can be found, finish the resolution process. Else, proceed to step 3.</Paragraph>
      <Paragraph position="6"> 3) Examine whether there are ,~ntecedents within other sentences in tit(.&amp;quot; text. (For example, a.naphora resolutim, is performed using Nakaiwa's method (Nakaiw~t and Ikeharlt, 1992)) If their ~mtecedents can be found, linish the resolution process. Else, proceed to step 4.</Paragraph>
      <Paragraph position="7"> 4) l)eictie resolution of Japanese zero pronouns using verb~tl semautic ~tttributes, modal expressions ;ted the types of conjunctious are conducted. The conditions to (let(:rminc tile referents are summ~trized in Table 2.</Paragraph>
      <Paragraph position="8"> If their referents ca.n be found, finish the resoh&gt; lion process. Else, proceed to step 5.</Paragraph>
    </Section>
    <Section position="7" start_page="814" end_page="815" type="sub_section">
      <SectionTitle>
4.1 Evaluation Method
</SectionTitle>
      <Paragraph position="0"> In this section, w('. show the results of evaluation of the method that was proposed above. The method to resolve zero pronouns with deictic ref- null The target was to resolve successfully the five types of zero pronouns (ga-case ~-- 'T' or &amp;quot;we&amp;quot;, ga-case e- &amp;quot;you&amp;quot; , ga-case +-- HUMAN, ga-case ~-- &amp;quot;it&amp;quot; , hi-case ~- &amp;quot;you&amp;quot;; 175 instances). These are the zero pronouns with deictic reference found within the 512 zero pronouns in the a718 sentence set for the evaluation of Japanese-to-English machine translation systems.</Paragraph>
      <Paragraph position="1">  The rules to resolve 175 zero pronouns were created by examining these zero pronouns using the constraints discussed in section 3 (46 rule@.</Paragraph>
      <Paragraph position="2">  'l'o examine the relationship between conditions of resolution and accuracy of resolution, we conducted the following two tests.</Paragraph>
      <Paragraph position="3"> (1) Resolution aeeuraey for eonditions of resolution We examined the accuracy of resolution depending on the types of conditions in a.naphora resolutiou such as semantic constraints to the cases, modal expression, verbal semantic attributes and conjunctive expressions. We evaluated the accuracy depending on the types of constraints used. (2) Resolution accuracy for rule complexity We examined the accuracy of the resohttions to see how they were affected by the complexities of the rules that were used in the resolution. In this test we evaluated the accuracy using simple, easily created and universal rules.</Paragraph>
    </Section>
    <Section position="8" start_page="815" end_page="815" type="sub_section">
      <SectionTitle>
4.2 Resolution Accuracy for Conditions
of Resolution
</SectionTitle>
      <Paragraph position="0"> To examine the resolution accuracy under different conditions, we examined the accuracy of the method proposed in this paper with the tbllowing 4 kinds of conditions: * using conditions of semantic constraints on cases only * using conditions of semantic constraints on cases and modal expression * using conditions of semantic constraints on cases, modal expression and verbal semantic attributes * using conditions of semantic constraints on cases, modal expression, verbal semantic attributes and conjunctions Table 3 shows the results of the resolution depending on the types of the rules. As shown in tAt the moment, it is difficult to use sentences which were not successfully syntactically and semantically a.nalyzed for the evaluation of our method. So, to evaluate the technical limitation of proposed method, we evaluated the resolution accuracy in the sentences which were examined to make the 46 rules (window test). We will conduct blind tests after we have finished debugging the whole system.</Paragraph>
      <Paragraph position="1"> this table, all 175 zero pronouns can be resolved using the rules that were proposed in section 3.</Paragraph>
      <Paragraph position="2"> The. introduction of verbal semantic attributes has achieved the same accuracy of resolution as the in-troduction of modM expressions (41 entries, 24%). From this result, we can say that the verbal semantic attributes are comparatively as effective as modal expressions. The results also show that , without using the constraints of conjunctions, the accuracy achieved is as high as 85%.</Paragraph>
    </Section>
    <Section position="9" start_page="815" end_page="815" type="sub_section">
      <SectionTitle>
4.3 Resolution Accuracy against Rule
Complexity
</SectionTitle>
      <Paragraph position="0"> To examine how the resolution accuracy varied according to the complexity of' rules, we tested the accuracy of the method proposed in this paper at different levels of complexity. The complexities C were evaluated using the following formula, and depended on the number of constraints used.</Paragraph>
      <Paragraph position="1"> C = # of modal const.* 1 + # of VSA coast.* 1 + # of conjunctions coast.* 2 In this formula, 1 in the modal and VSA and 2 in the conjunction indicate the weights. Because conjunction constraints affect both sides of the unit sentence, we gave the conjunctions constraints a weight of 2. According to this formula, the complexity of a rule that has a constraint fbr conjunctions and for VSA in the main clause and for modM and VSA in the sub clause, becoines 5(-l(modal)*l + I(VSA)*2 + 2(conjunction)*l).</Paragraph>
      <Paragraph position="2"> Table 4 shows the accuracy of the resolution depending on the complexities of the rules. 46 kinds of rules were used in the deictic resolution of 175 zero pronouns as shown in |;able 4. The accuracy of resolution using rules with complexities of 3 or less, is 90%, and the accuracy of resolution using rules with complexities of 4 or less, is 95%.</Paragraph>
      <Paragraph position="3"> This result shows that the use of the constraints based on modM expressions, VSA and conjunctions can achieve high accuracy using relatively simple rules.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>