Towards a Representation of Idioms in WordNet

1 I n t r o d u c t i o n WordNet (Miller, 1995), (Fellbaum, 1998) is perhaps the most widely used electronic dictionary of English and serves as the lexicon for a rarity of different NLP applications including Information Retrieval (IR), Word Sense Disambignation (WSD), and M~hine Transla~ tion (MT). Despite WordNet's large coverage, which comprises some 100,000 concepts lexicMi~.ed by approySmately 120,000 word forms (strings) and is comparable to that of a collegiate dictionary, it contains relatively little figurative language. WordNet includes a w~mber of multi-word strings, such as phrasal verbs, but many idiomatic verb phrases Like smell a rat, know the ropes, and eat humble pie, are mi~ging. Idioms and metaphors abound in everyday language and are found in texts spanning many genres (see, e.g., (Jackendoff, 1997) for a numerical estlm~te of the frequency of idioms and fixed expression). Clearly, a dictionary that indudes extended senses of words and phrases is likely to yield more successful NLP applications. On the one hand, no system wants to retrieve the string bucket from the idiom kick the bucket. On the other hand, MT and WSD efforts need to distinguish the sense of ropes in phrases like know~learn/teach someone the ropes from the sense meaning "strong cords"; selecting the latter sense in any of the idiomatic phrases leads to failure. An IR query is likely to be interested only in the "strong cord" reading. When this sense is to be retrieved with the aid of a lexicon intended for multiple applications, the figurative sense must be successfully recognized and excluded from a text that may contain instances of the string ropes with both meanings. In this paper, we consider the possibility of extending WordNet to accommodate figurative meanings in the English lexicon. While much