Reusing an ontology to generate numeral classifiers

Reusing an ontology to generate numeral classifiers Francis Bond  NTT Communication Science Laboratories 2-4 Hikari-dai, Kyoto 619-0237, JAPAN bond@cslab.kecl.ntt.co.jp Kyonghee Paik Center for the Study of Language and Information Stanford University, CA 94305-2150, USA kpaik@usa.net Abstract In this paper, we present a solution to the prob- lem of generating Japanese numeral classifiers us- ing semantic classes from an ontology. Most nouns must take a numeral classifier when they are quan- tified in languages such as Chinese, Japanese, Ko- rean, Malay and Thai. In order to select an appro- priate classifier, we propose an algorithm which as- sociates classifiers with semantic classes and uses inheritance to list only those classifiers which have to be listed. It generates sortal classifiers with an ac- curacy of 81%. We reuse the ontology provided by Goi-Taikei — a Japanese lexicon, and show that it is a reasonable choice for this task, requiring infor- mation to be entered for less than 6% of individual nouns. 1 Introduction In this paper we consider two questions. The first is: how to generate numeral classifiers such as piece in 2 pieces of paper? To do this we use a semantic hier- archy originally developed for a different task. The second is: how far can such a hierarchy be reused? In English, uncountable nouns cannot be directly modified by numerals, instead the noun must be embedded in a noun phrase headed by a classi- fier. Knowing when to do this is a language spe- cific property. For example, French deux renseigne- ment must be translated as two pieces of information in English.1 In many languages, including most South-East Asian languages, Chinese, Japanese and Korean, the majority of nouns are uncountable and must be quantified by numeral classifier combina- tions. These languages typically have many differ- ent classifiers. There has been some work on the analysis of numeral classifiers in natural language processing, particularly for Japanese (Asahioka et al., 1990; Kamei and Muraki, 1995; Bond et al.,  Visiting CSLI, Stanford University (1999-2000). 1Numeral-classifier combinations are shown in bold, the noun phrases they quantify are underlined. 1996; Bond et al., 1998; Yokoyama and Ochiai, 1999), but very little on their generation. We could only find one paper on generating classifiers in Thai (Sornlertlamvanich et al., 1994). One immediate application for the generation of classifiers is ma- chine translation, and we shall take examples from there, but it is in fact needed for the generation of any quantified noun phrase with an uncountable head noun. The second question we address is: how far can an ontology be reused for a different task to the one it was originally designed for. There are several large ontologies now in use (WordNet (Fellbaum, 1998); Goi-Taikei (Ikehara et al., 1997); Mikrokos- mos (Nirenburg, 1989)) and it is impractical to re- build one for every application. However, there is no guarantee that an ontology built for one task will be useful for another. The paper is structured as follows. In Section 2, we discuss the properties …

Content truncated to 500 words