Generating “A for Alpha” When There Are Thousands of Characters

Proceedings of COLING 2012: Technical Papers, pages 1411–1424, COLING 2012, Mumbai, December 2012. Generating “A for Alpha” When There Are Thousands of Characters Hiroaki KAWASAKI1 Ryohei SASANO2 Hiro ya TAKAMURA2 Manabu OKUMURA2 (1) Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology (2) Precision and Intelligence Laboratory, Tokyo Institute of Technology kawa@lr.pi.titech.ac.jp, {sasano,takamura,oku}@pi.titech.ac.jp ABSTRACT The phonetic alphabet enables people to dictate letters of the alphabet accurately by using representative words, i.e., A for Alpha. Japanese kanji (idiographic Chinese characters) vastly outnumber the letters of the Roman alphabet, and thus Japanese requires an explanatory reading like a phonetic alphabet. We call the explanatory reading of a kanji a “distinctive explanation.” Most kanji characters have their homophones, and the role of the distinctive explanations is to enable users to identify a specific kanji character only by listening to the explanation. In this paper, we propose a corpus-based method for automatically generating distinctive explanations for a kanji, in which information about familiarity and homophones of kanji are taken into consideration. Through the kanji-identification experiments, we show that the quality of the explanations generated by the proposed method is higher than that of the manually crafted distinctive explanations. KEYWORDS: phonetic alphabet, distinctive explanation for a kanji. 1411 1 Introduction Japanese has three types of characters: hiragana, katakana, and kanji (Chinese characters). While hiragana and katakana characters are phonograms, kanjis are ideograms, each of which usually has several readings. Most kanjis have homophones, and thus it is difficult to identify a kanji only by its reading. However, sometimes we need to identify a kanji verbally. For example, screen readers need to enable the users, especially visually impaired people who have difficulty seeing things, to identify a kanji only by sound. When we talk over the telephone, we have to exchange information, such as proper names, only by our voice. In such cases, we explain a kanji by using not only its reading but also its properties, compositions, and so on, to reduce the ambiguity. In this paper, we call such an explanation a “distinctive explanation for a kanji” and propose a method for generating distinctive explanations automatically. One concept similar to a distinctive explanation is the phonetic alphabet, e.g., A for Alpha, B for Bravo, and C for Charlie. While a distinctive explanation explains kanjis, the phonetic alphabet explains letters and numbers. One of the major differences between them is the number of target characters. Whereas the phonetic alphabet for English deals with only 26 letters and 10 numbers, the distinctive explanation for kanjis deals with thousands of kanji characters. The distinctive explanation leverages various aspects of the target kanji such as its Chinese reading, its Japanese reading, and its radicals, to reduce ambiguity. Each kanji has several readings, some of which are derived from Chinese readings and the others are Japanese tra- ditional readings. If a kanji has a distinctive reading, the reading can be used to generate the distinctive explanation. Some kanjis can be divided into left and right parts and …

Content truncated to 500 words