File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/03/p03-2013_metho.xml
Size: 19,862 bytes
Last Modified: 2025-10-06 14:08:19
<?xml version="1.0" standalone="yes"?> <Paper uid="P03-2013"> <Title>Approaches to Zero Adnominal Recognition</Title> <Section position="3" start_page="2" end_page="2" type="metho"> <SectionTitle> 2 Zero Adnominals </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 2.1 Definition </SectionTitle> <Paragraph position="0"> Recall the discourse segment in (1). Its original Japanese is analyzed in (2).</Paragraph> <Paragraph position="1"> (2) a. simauma-wa raion ni itumo zebra-TOP lion-DAT always ki-o-tuke-nakereba-narimasen.</Paragraph> <Paragraph position="2"> watch-out-for-need-to &quot;Zebras always need to watch out for lions.&quot; b. desukara, O kusa-o tabete-ite-mo, so O-NOM grass-ACC eating-even-while &quot;So even while (they) are eating grass,&quot; c. O O usiro-no-ho-made mieru-yo-ni O-NOM O-ADN-behind-even see-can-for &quot;so that (they) can see even what is behind (them),&quot; d. O me-ga O kao-no-yoko-ni O-ADN-eye-NOM O-ADN-face-side LOC tuite-imasu.</Paragraph> <Paragraph position="3"> placed-be &quot;(their)eyes are on the sides of (their) faces.&quot; Zero arguments are unexpressed elements that are predictable from the valency requirements of their heads, i.e., a given predicate of the clause. Zero nominatives in (2b) and (2c) are of this type. Zero adnominals, analogously, are missing elements that can be inferred from some features specified by their head nouns. A noun for body-part, me 'eyes' in (2d) usually calls hearers' attention to &quot;ofwhom&quot; information and hearers recover that information in the flow of discourse. That missing information can be supplied by a noun phrase (NP) followed by an adnominal particle no, i.e., simauma-no 'zebras'(= their)' in the case of (2d) above. Hence, as a first approximation, we define a zero adnominal as an unexpressed &quot;NP no&quot; in the NP no NP (a.k.a., A no B) construction.</Paragraph> </Section> <Section position="2" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 2.2 The Corpus </SectionTitle> <Paragraph position="0"> Before we proceed, we will briefly describe the corpus that we investigated. The corpus consists of a collection of 83 written narrative texts taken from seven different JSL textbooks with levels ranging from beginning to intermediate. Thus, it is a representative sample of naturally-occurring, but maximally canonical, free-from-deviation, and coherent narrative discourse.</Paragraph> </Section> <Section position="3" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 2.3 Identification </SectionTitle> <Paragraph position="0"> Our primary goal is to identify relevant information for recognizing zero adnominals. Since such information is unavailable in the surface text, the identification of missing adnominal elements and their referents in the corpus was based on the native speaker intuitions and the linguistic expertise of the author, who used the definition in 2.1, with occasional consultation with a JSL teaching expert/linguist. As a result, we located a total of 320 zero adnominals. These adnominals serve as the zero adnominal samples on which our later analysis is based.</Paragraph> </Section> </Section> <Section position="4" start_page="2" end_page="2" type="metho"> <SectionTitle> 3 Theoretical/Pedagogical Motivations </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 3.1 Centering Analysis </SectionTitle> <Paragraph position="0"> One discourse account that models the perceived degree of coherence of a given discourse in relation to local focus of attention and the choice of referring expressions is centering (e.g., Grosz, Joshi and Weinstein, 1995).</Paragraph> <Paragraph position="1"> The investigation of zeros behavior in our corpus, within the centering framework, shows that zero adnominals make a considerable contribution to center continuity in discourse by realizing the central entity in an utterance (called Cb) just as well-acknowledged zero arguments do.</Paragraph> <Paragraph position="2"> Recall example (2). Its center data structure is given in (3). The Cf (forward-looking center) list is a set of discourse entities that appear in each</Paragraph> <Paragraph position="4"> ). The Cb (backward-looking center) is a special member of the Cf list, and is meant to represent the entity that the utterance is most centrally about; it is the most highly ranked element of</Paragraph> <Paragraph position="6"> (3) a. Cb: none [Cf: zebra, lion] b. Cb: zebra [Cf: zebra, grass] c. Cb: zebra [Cf: zebra, what is behind] d. Cb: zebra [Cf: zebra, eye, face-side] In (3b) and (3c), the Cb is realized as a zero nominative, and in (3d), it is realized by the same entity (zebra) as a zero adnominal, maintaining the CONTINUE transition that by definition is maximally coherent. This matches the intuitively perceived degree of coherence in the utterance. Our corpus contains a total of 138 zero adnominals that refer to previously mentioned entities (15.56% of all the zero Cbs), and realize the Cb of the utterance in which they occur, as in (3d=2d).</Paragraph> <Paragraph position="7"> Our corpus study shows that discourse coherence can be more accurately characterized, in the centering account, by recognizing the role of zero adnominals as a valid realization of Cbs (see Yamura-Takei et al., ms. for detailed discussion). This is our first motivation towards zero adnominal recognition.</Paragraph> </Section> <Section position="2" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 3.2 Zero Detector </SectionTitle> <Paragraph position="0"> Yamura-Takei et al. (2002) developed an automatic zero identifying tool. This program, Zero Detector (henceforth, ZD) takes Japanese written narrative texts as input and provides the zerospecified texts and their underlying structures as output. This aims to draw learners' and teachers' attention to zeros, on the basis of a hypothesis about ideal conditions for second language acquisition, by making invisible zeros visible. ZD regards teachers as its primary users, and helps them predict the difficulties with zeros that students might encounter, by analyzing text in advance. Such difficulties often involve failure to recognize discourse coherence created by invisible referential devices, i.e., the center continuity maintained by the use of various types of zeros.</Paragraph> <Paragraph position="1"> As our centering analysis above indicates, inclusion of zero adnominals into ZD's detecting capability enables a more comprehensive coverage of the zeros that contributes to discourse coherence.</Paragraph> <Paragraph position="2"> This is our project goal.</Paragraph> </Section> </Section> <Section position="5" start_page="2" end_page="2" type="metho"> <SectionTitle> 4 Towards Zero Adnominal Recognition </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="2" end_page="2" type="sub_section"> <SectionTitle> 4.1 Semantic Classification </SectionTitle> <Paragraph position="0"> Unexpressed elements need to be predicted from other expressed elements. Thus, we need to characterize B nouns (which are overt) in the (A no) B construction, assuming that zero adnominals (A) are triggered by their head nouns (B) and that certain types of NPs tend to take implicit (A) arguments. Our first approach is to use an existing A no B classification scheme. We adopted, from among many A no B works, a classification modeled on Shimazu, Naito and Nomura (1985, 1986, and 1987) because it offers the most comprehensive classification (Fais and Yamura-Takei, ms). Table 1 below describes the five main groups that we used to categorize (A no) B phrases.</Paragraph> </Section> </Section> <Section position="6" start_page="2" end_page="7" type="metho"> <SectionTitle> 4.2 Results </SectionTitle> <Paragraph position="0"> We classified our 320 &quot;(A no) B&quot; examples into the five groups described in the previous section.</Paragraph> <Paragraph position="1"> Group V comprised the vast majority, while approximately the same percentage of examples was included in Groups I, II and III. There were no Group IV examples. The number and percentage of examples of each group are presented in Table 2.</Paragraph> <Paragraph position="2"> We conjecture that certain nouns are more likely to take zero adnominals than others, and that the head nouns which take zero adnominals, extracted from our corpus, are representative samples of this particular group of nouns. We call them &quot;argument-taking nouns (ATNs).&quot; ATNs syntactically require arguments and are semantically dependent on their arguments. We use the term ATN only to refer to a particular group of nouns that can take implicit arguments (i.e., zero adnominals).</Paragraph> <Paragraph position="3"> We closely examined the 127 different ATN tokens among the 320 cases of zero adnominals and classified them into the four types that correspond to Groups I, II, III and V in Table 1. We then listed their syntactic/semantic properties based on the syntactic/semantic properties presented in the Goi-Taikei Japanese Lexicon (hereafter GT, Ikehara, Miyazaki, Shirai, Yokoo, Nakaiwa, Ogura, Oyama, and Hayashi, 1997). GT is a semantic feature dictionary that defines 300,000 nouns based on an ontological hierarchy of approximately 2,800 semantic attributes. It also uses nine part-of-speech codes for nouns. Table 3 lists the syntactic/semantic characterizations of the nouns in each type and the number of examples in the corpus. What bold means in the table will be When we examine these four types, we see that they partially overlap with some particular types of nouns studied theoretically in the literature. Teramura (1991) subcategorizes locative relational nouns like mae 'front', naka 'inside', and migi 'right' as &quot;incomplete nouns&quot; that require elements to complete their meanings; these are a subset of Type II. Iori (1997) argues that certain nouns are categorized as &quot;one-place nouns,&quot; in which he seems to include Type I and some of Type V nouns.</Paragraph> <Paragraph position="4"> Kojima (1992) examines so-called &quot;lowindependence nouns&quot; and categorizes them into three types, according to their syntactic behaviors in Japanese copula expressions. These cover sub-sets of our Type I, II, III and V. In computational work, Bond, Ogura, and Ikehara (1995) extracted 205 &quot;trigger nouns&quot; from a corpus aligned with English. These nouns trigger the use of possessive pronouns when they are machine-translated into English. They seem to correspond mostly to our Type V nouns. Our result offers a comprehensive coverage which subsumes all of the types of nouns discussed in these accounts.</Paragraph> <Paragraph position="5"> Next, let us more closely look at the properties expressed by our samples. The most prevalent ATNs (21 in number) are nominalized verbals in the semantic category of human activity. The next most common are kinship nouns (14 in number) and body-part nouns (14), both in the common noun category; location nouns (13), either in the common noun or formal noun category; and nouns that express amount (9) whose syntactic category is either common or de-adjectival. The others include some &quot;human&quot; subcategories, etc.</Paragraph> <Paragraph position="6"> The part-of-speech subcategory, &quot;nominalized verbal&quot; (sahen-meishi) is a reasonably accurate indicator of Type 1 nouns. So is &quot;formal noun&quot; (keishiki-meishi) for Type II, although this does not offer a full coverage of this type. Numeral noun and counter suffix noun compounds also represent a major subset of Type III.</Paragraph> <Paragraph position="7"> Semantic properties, on the other hand, seem helpful to extract certain groups such as location (Type II), amount (Type III), kinship, body-part, organization, and some human subcategories (Type V). But other low-frequency ATN samples are problematic for determining an appropriate level of categorization in GT's semantic hierarchy tree.</Paragraph> <Section position="1" start_page="2" end_page="4" type="sub_section"> <SectionTitle> 4.3 Algorithm </SectionTitle> <Paragraph position="0"> Our goal is to build a system that can identify the presence of zero adnominals. In this section, we propose an ATN (hence zero adnominal) recognition algorithm. The algorithm consists of a set of lexicon-based heuristics, drawn from the observations in section 4.2.</Paragraph> <Paragraph position="1"> The algorithm takes morphologically-analyzed text as input and provides ATN candidates as output. The process consists of the following three phases: (i) bare noun extraction, (ii) syntactic category (part-of-speech) checking, and (iii) semantic category checking.</Paragraph> <Paragraph position="2"> Zero adnominals usually co-occur with &quot;bare nouns.&quot; Bare nouns, in our definition, are nouns without any pre-nominal modifiers, including demonstratives, explicit adnominal phrases, relative clauses, and adjectives.</Paragraph> <Paragraph position="3"> Bare nouns are often simplex as in (4a), and sometimes are compound (e.g., numeral noun + counter suffix noun) as in (4b). These are immediately followed by case-marking, topic/focus-marking or other particles (e.g., ga, o, ni, wa, mo).</Paragraph> <Paragraph position="4"> (4) a. atama-ga head-NOM b. 70-paasento-o 70-percent-ACC The extracted nouns under this definition are initial candidates for ATNs.</Paragraph> <Paragraph position="5"> Once bare nouns are identified, they are checked against our syntactic-property- (i.e., partof-speech, POS) based-, followed by semantic-attribute (SEM) based-heuristics. For semantic filtering, we decided to use the noun groups of high frequency (more than two tokens categorized in the same group; indicated in bold in Table 3 above) to minimize a risk of over-generalization. The algorithm checks the following two conditions, for each bare noun, in this order: [1] If POS = [nominalized verval, derived noun, formal noun, numeral + counter suffix compound], label it as ATN.</Paragraph> <Paragraph position="6"> [2] If SEM = [2610: location, 2585: amount, 362: organization, 552: animate (part), 111: human (relation), 224: human (profession), 72: Japanese do not use determiners for its nouns. human (kinship), 866: housing (part), 813: clothing], label it as ATN.</Paragraph> <Paragraph position="7"> Therefore, nouns that pass condition [1] are labeled as ATNs, without checking their semantic properties. A noun that fails to pass condition [1] and passes condition [2] is labeled as ATN. A noun that fails to match both [1] and [2] is labeled as non-ATN. Consider the noun sintyo 'height' for example. Its POS code in GT is common noun, so it fails condition [1] and goes to [2]. This noun is categorized in the &quot;2591: measures&quot; group which is under the &quot;2585: amount&quot; node in the hierarchy tree, so it is labeled as ATN. In this way, the algorithm labels each bare noun as either ATN or nonATN. null</Paragraph> </Section> <Section position="2" start_page="4" end_page="7" type="sub_section"> <SectionTitle> 4.4 Evaluation </SectionTitle> <Paragraph position="0"> To assess the performance of our algorithm, we ran it by hand on a sample text.</Paragraph> <Paragraph position="1"> The test corpus contains a total of 136 bare nouns. We then matched the result against our manually-extracted ATNs (34 in number). The result is shown in Table 4 below, with recall and precision metrics. As a baseline measurement, we give the accuracy for classifying every bare noun as ATN. For comparison, we also provide the results when only either POS-based or semantic-based heuristics are applied.</Paragraph> <Paragraph position="2"> Semantic categories make a greater contribution to identifying ATNs than POS. However, the POS/Semantic algorithm achieved a higher recall but a lower precision than the semantic-only algorithm did. This is mainly because the former produced more over-detected errors. Closer examination of those errors indicates that most of them (8 out of 9 cases) involve verbal idiomatic expressions that contain ATN candidate nouns, as example (5) shows.</Paragraph> <Paragraph position="3"> These numbers indicate the numbers assigned to each semantic category in Goi-Taikei Japanese Lexicon (GT). This is taken from the same genre as our corpus for the initial analysis, i.e., another JSL textbook.</Paragraph> <Paragraph position="4"> (5) me-o-samasu eye-ACC-wake 'wake up' Although me 'eye' is a strong ATN candidate, as in example (2) above, case (5) should be treated as part of an idiomatic expression rather than as a zero adnominal expression.</Paragraph> <Paragraph position="5"> Thus, we decided to add another condition, [0] below, before we apply the POS/SEM checks. The revised algorithm is as follows: [0] If part of idiom in [idiom list], label it as non-ATN.</Paragraph> <Paragraph position="6"> [1] If POS = [nominalized verval, derived noun, formal noun, numeral + counter suffix compound], label it as ATN.</Paragraph> <Paragraph position="7"> [2] If SEM = [2610: location, 2585: amount, 362: organization, 552: animate (part), 111: human (relation), 224: human (profession), 72: human (kinship), 866: housing (part), 813: clothing], label it as ATN.</Paragraph> <Paragraph position="8"> When a noun matches condition [0], it will not be checked against [1] and [2]. When this applies, the evaluation result is now as shown below. The revised algorithm, with both syntactic/semantic heuristics and the additional idiom-filtering rule, achieved a precision of 96.96%. The result still includes some over/under-detecting errors, which will require future attention.</Paragraph> </Section> </Section> <Section position="7" start_page="7" end_page="8" type="metho"> <SectionTitle> 5 Related Work </SectionTitle> <Paragraph position="0"> Associative anaphora (e.g., Poesio and Vieira, 1998) and indirect anaphora (e.g., Murata and Nagao, 2000) are virtually the same phenomena that this paper is concerned with, as illustrated in (6).</Paragraph> <Paragraph position="1"> Vieira and Poesio (2000) also list &quot;idiom&quot; as one use of definite descriptions (English equivalent to Japanese bare nouns), along with same head/associative anaphora, etc.</Paragraph> <Paragraph position="2"> The list currently includes eight idiomatic samples from the test data, but it should of course be expanded in the future. (6) a. a house - the roof b. ie 'house' - yane 'roof' c. ie 'house' - (O-no) yane '(O's) roof' We take a zero adnominal approach, as in (6c), because we assume, for our pedagogical purpose discussed in section 3.2, that zero adnominals, by making them visible, more effectively prompt people to notice referential links than lexical relations, such as meronymy in (6a) and (6b).</Paragraph> <Paragraph position="3"> However, insights from other approaches are worth attention. There is a strong resemblance between bare nouns (that zero adnominals co-occur with) in Japanese and definite descriptions in English in their behaviors, especially in their referential properties (Sakahara, 2000). The task of classifying several different uses of definite descriptions (Vieira and Poesio, 2000; Bean and Riloff, 1999) is somewhat analogous to that for bare nouns. Determining definiteness of Japanese noun phrases (Heine, 1998; Bond et al., 1995; Murata and Nagao, 1993) is also relevant to ATN (which is definite in nature) recognition.</Paragraph> </Section> class="xml-element"></Paper>