<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-1407">
  <Title>Naruaki Masuno ++</Title>
  <Section position="4" start_page="1" end_page="41" type="metho">
    <SectionTitle>
2 Preliminary investigation
</SectionTitle>
    <Paragraph position="0"> To make an investigation into the variation and distribution of required verbal suffixes, we collected a set of paraphrase examples through the following semi-automatic procedure: Step 1. We handcrafted adjective-verb pairs based on JCore (Sato, 2004), which classifies Japanese words into five-levels of readability.</Paragraph>
    <Paragraph position="1"> Our 128 pairs (for 85 adjectives) contain only those sharing first few phonemes (reading)  For each example, &amp;quot;s&amp;quot; and &amp;quot;t&amp;quot; denote an original sentence and its paraphrase, respectively.</Paragraph>
    <Paragraph position="3"> re-ta 6 0 1 re-tei-ta 0 0 1 both ta and tei-ru 4 0 0 both ta and ru 1 0 0</Paragraph>
    <Paragraph position="5"> tei: progressive / perfective re: passive / potential ta: past / attributive tea: perfective and characters (kanji), and either of adjective or verb falls into the easiest three levels. Step 2. Candidate paraphrases for a given sentence collection are automatically generated by replacing adjectives with their corresponding verbs. Multiple candidates are generated for adjectives that correspond to multiple verbs.</Paragraph>
    <Paragraph position="6"> Step 3. The correctness of each candidate paraphrase is judged by two human annotators. The basic criterion for judgement is that two sentences are regarded as paraphrases if and only if they share at least one interpretation. In this step, the annotators are allowed to revise candidates: (i) append verbal suffixes, (ii) change of case markers, and (iii) insert adverbs. Finally, candidates that both annotators judge correct qualify as paraphrases. Assuming that the variation and distribution of verbal suffixes vary according to the usage of adjectives, we separately collected paraphrase examples for adnominal and predicative usages. Adnominal usages: For 960 sentences randomly extracted from a one-year newspaper corpus, Mainichi 1995, we obtained 165 examples for 142 source sentences. We then divided them into two portions: 12 adjectives that appeared only once and at least one examples for the other adjectives were kept unseen (C</Paragraph>
    <Paragraph position="8"> ), while the remaining ex-</Paragraph>
    <Paragraph position="10"> ) were used for our investigation.</Paragraph>
    <Paragraph position="11"> Predicative usages: For 157 example sentences within IPAL adjective dictionary (IPA, 1990), we generated candidate paraphrases. 84 candidates for 70 sentences qualified as paraphrases. They are then divided into two portions according to the tense of adjectives: C</Paragraph>
    <Paragraph position="13"> consists of examples where adjectives appear in base form and C</Paragraph>
    <Paragraph position="15"> for &amp;quot;ta&amp;quot; form (past tense).</Paragraph>
    <Paragraph position="16"> Table 1 shows the distribution of verbal suffixes used for given adjective-verb pairs in each portion of example collections. We confirmed that their distribution was fairly different. In the remaining sections, we focus on adnominal usages because examples of predicative usages have displayed a degree of compositionality. Which of &amp;quot;ru&amp;quot;or&amp;quot;ta&amp;quot; must be used is given by the input: if a given adjective accompanies past tense, the resultant verbal suffix is necessarily that for present tense followed by &amp;quot;ta.&amp;quot;</Paragraph>
  </Section>
  <Section position="5" start_page="41" end_page="42" type="metho">
    <SectionTitle>
3 Determining verbal suffixes
</SectionTitle>
    <Paragraph position="0"> The task we address here is to determine verbal suffixes for a given input, a pair of an adnominal usage of adjective in a certain context and a candidate verb given by our adjective-verb list.</Paragraph>
    <Paragraph position="1"> From the viewpoint of language generation, this task can be thought of as generating verbal expressions where options are already given in  verbal suffixes is to make use of lexical properties of verbs as constraints on generation. To manifest them, in particular aspectual properties involved in LCS, we first designed seven types of linguistic tests shown in Table 2. They are derived from a classical analysis of verb semantics in Japanese (Kageyama, 1996) and some ongoing projects on constructing LCS dictionaries (Kato et al., 2005; Takeuchi et al., 2006). We then manually examined 128 verbs in Section 2 under those tests. To determine the word sense in which the derivative relationship hold good, example sentences in IPAL verb dictionary (IPA, 1987) for each verb were used. For a verb which was out of the dictionary, we manually gave a sample sentence.</Paragraph>
    <Paragraph position="2"> Since our aim is to explain why a certain verbal suffix is used for a given input, we have not feverishly applied a machine learning algorithm to the task. Instead, we have manually created a rule-based model shown in Table 3 using C</Paragraph>
    <Paragraph position="4"> * D: affix pair of the adjective and the candidate verb: e.g., &amp;quot;A shii-V mu&amp;quot; for &amp;quot;kuyashii (be regretful)&amp;quot; = &amp;quot;kuyamu (to regret)&amp;quot; * N: disjunction of semantic classes in a thesaurus (The Natural Institute for Japanese Language, 2004) for the modified noun * C: whether the adjective is head of clause 4 Experiment and discussion By conducting an empirical experiment with C</Paragraph>
    <Paragraph position="6"> , we evaluate how our model (RULE) properly determines verbal suffixes. A comparison with a simple baseline model (BL)isalso done. BL selects the most frequently used suffix (in this experiment &amp;quot;ta&amp;quot;) for any given input.  Table 4 shows the experimental results, where recall and precision are calculated with regard to input adjective-verb pairs. Among rules in Table 3, rules 1 (for &amp;quot;re-ru&amp;quot;),3,6,and7(for&amp;quot;ta&amp;quot; where V a =&amp;quot;no&amp;quot;) performed much better than the other rules. This indicates that these rules and features in their conditions properly reflect our linguistic intuition. For instance, rule 6 reflects that a change-of-state intransitive verb expresses resultative meaning as adjectives when it modifies Theme of the event via &amp;quot;ta&amp;quot; (Kageyama, 1996) as shown in (1), and rule 2 does that a psychological verb modifies a nouns with &amp;quot;re-ru&amp;quot;when the noun arouses the specific emotion, such as regretting mistakes (e.g., &amp;quot;kuyashii (be regretful)&amp;quot; = &amp;quot;kuyama-re-ru (be regretted)&amp;quot;). The aspectual property captured by the tests in Table 2 is used to classify verbs into these semantic classes.</Paragraph>
    <Paragraph position="7"> On the other hand, the rules for the other types are immature due to lack of examples: we cannot find out even necessary conditions to be &amp;quot;ru,&amp;quot; &amp;quot;teiru,&amp;quot; etc. What is required to induce proper conditions for these suffixes is a larger example collection and discovering another semantic property and a set of linguistic tests for capturing it.</Paragraph>
  </Section>
class="xml-element"></Paper>