Towards a Representation of Verbal Semantics – An Approach Based on Near-Synonyms

Computatioal Linguistics and Chinese Language Processing vol 3, no. 1 , February 98, pp. 61-74 �Computational Linguisitcs Society of R. O. C. Towards a Representation of Verbal Semantics -- An Approach Based on Near-Synonyms Mei-Chih Tsai*, Chu-Ren Huang*, Keh-Jiann Chen*, Kathleen Ahrens+ Abstract In this paper we propose using the distributional differences in the syntactic patterns of near-synonyms to deduce the relevant components of verb meaning. Our method involves determining the distributional differences in syntactic patterns, deducing the semantic features from the syntactic phenomena, and testing the semantic features in new syntactic frames. We determine the distributional differences in syntactic patterns through the following five steps: First, we search for all instances of the verb in the corpus. Second, we classify each of these instances into its type of syntactic function. Third, we classify each of these instances into its argument structure type. Fourth, we determine the aspectual type that is associated with each verb. Lastly, we determine each verb's sentential type. Once the distributional differences have been determined, then the relevant semantic features are postulated. Our goal is to tease out the lexical semantic features as the explanation, and as the motivation of the syntactic contrasts. 1. Introduction Radical Lexicalism maintains that all grammatical behaviors are manifestations of lexical features [Karttunen 1986]. Since most lexical attributes are semantic and/or conceptual in nature, taking this lexicon-driven approach to language means that many syntactic properties can be predicted from lexical semantic attributes [Jackendoff 1976, Levin 1985, Dowty 1991, Pustejovsky 1993]. In terms of Natural Language Processing (NLP), surface syntactic structures can be systematically predicted from their lexical semantic representation. From this perspective, the automatic acquisition of lexical knowledge for NLP may be possible, since the relation between syntactic patterns and lexical semantics is predictable to some extent. Dorr & Jones [1996], for example, demonstrate that semantic information can be derived from syntactic cues when the syntactic cues are first divided into distinct groupings that correlate with different word senses. * Academia Sinica, Nankang, Taipei 115, Taiwan, R.O.C. E-mail: tsmei@mail.nsysu.edu.tw + National Taiwan University, Taipei 106, Taiwan, R.O.C. 61 However, as Levin [1993] points out, there are still many questions to be explored: If the hypothesis that syntactic properties are semantically deter- mined is taken seriously, then the task is to determine, first, to what extent the meaning of a verb determines its syntactic behavior, and second, to the extent that syntactic behavior is predictable, what components of verb meaning figure in the relevant generalizations. The identification of the relevant components of meaning is essential if this approach is to be successful.[Levin 1993:14] Our paper will focus on the last point above. We propose using the distributional differences in the syntactic patterns of near-synonyms1 to deduce the relevant compo- nents of verbal semantics. In particular, we want to identify the semantic features that differentiate verbal syntactic behaviors. Our strong hypothesis is that all lexical semantic features can be identified this way. In contrast, salient semantic features deduced from a shared verb class …

Content truncated to 500 words