Metonymy Analysis Using Associative Relations between Words

Metonymy Analysis Using Associative Relations between Words Takehiro Teraoka School of Media Science, Tokyo University of Technology 1404-1, Katakura, Hachioji, Tokyo, 192-0914 Japan teraokatkhr@stf.teu.ac.jp Abstract Metonymy is a figure of speech in which one item’s name represents another item that usually has a close relation with the first one. Metonymic expressions need to be correctly detected and interpreted because sentences including such expressions have different mean- ings from literal ones; computer systems may output inappropriate results in natural language processing. In this paper, an associative approach for analyzing metonymic expressions is proposed. By using associative information and two conceptual distances between words in a sentence, a previous method is enhanced and a decision tree is trained to detect metonymic expressions. After detecting these expressions, they are interpreted as metonymic understanding words by using associative information. This method was evaluated by comparing it with two baseline methods based on previous studies on the Japanese language that used case frames and co-occurrence information. As a result, the proposed method exhibited significantly better accuracy (0.85) of determining words as metonymic or literal expressions than the baselines. It also exhibited better accuracy (0.74) of interpreting the detected metonymic expressions than the baselines. Keywords: Metonymy, Associative Concept Dictionary, Verb, Noun 1. Introduction Metonymy is a figure of speech, in which one item’s name represents another item that usually has a close relation to the first. The relations shown in Table 1 (Lakoff and John- son, 1980; Yamanashi, 1988) have various patterns based on contiguity. Below is a Japanese metonymic example of the spatial contiguity for ‘Container for Content’ and its di- rect English translation: kare-ga isshoubin-wo nomihoshita ‘He drank a 1.8-liter bottle.’ The Japanese sentence does not mean that he drank the bot- tle itself, but its contents, usually Japanese sake. Japanese sake is generally served from a 1.8-liter glass bottle, called isshoubin. Therefore, the above example sentence, where isshoubin is a metonymic expression, means that he drank Japanese sake from a 1.8-liter bottle. Since a sentence in- cluding metonymy is grammatical on a literal level and metonymic relations have many patterns, it is difficult for computers to analyze these expressions. Previously, rule-based or statistical approaches have been used to detect and interpret metonymy. Rule-based ap- proaches use semantic networks and rules to discriminate metonymy (Bouaud et al., 1996; Fass, 1991). Statistical approaches use corpus-based metonymy resolution on lo- cation names, which is done with machine learning tech- niques (Markert and Nissim, 2003). Moreover, by using syntactic or semantic information as machine learning fea- tures, five different studies classified the large data set of metonymic and literal meanings in the SemEval-2007 Task 08: Metonymy Resolution at SemEval-2007 (Nicolae et al., 2007; Poibeau, 2007). Previous studies in the Japanese lan- guage also used thesauri to detect metonymy and statistical approaches with example-based information extracted from newspaper corpora in interpreting metonymy (Murata et al., 2000). Metonymic patterns Examples of sentences -spatial contiguity- (metonymic reading) Container for Content kare-ha glass-wo nonda ‘He drank the glass (liquid).’ Producer for …

Content truncated to 500 words