Ten Choices for Lexical Semantics

The modern computational lexical semantics reached a point in its development when it has become necessary to define the premises and goals of each of its several trends. This paper proposes ten choices in terms of which these premises and goals can be discussed. It is argued that the central questions include the use of lexical rules for generating word senses; the role of syntax, pragmatics, and formal semantics in the specification of lexical meaning; the use of a world model, or ontology, as the organizing principle for lexical-semantic descriptions; the use of rules with limited scope; the relation between static and dynamic resources; the commitment to descriptive coverage; the trade-off between generalization and idiosyncracy; and, finally, the adherence to the “supply-side” (method-oriented) or “demand-side” (task-oriented) ideology of research. The discussion is inspired by, but not limited to, the comparison between the generative lexicon approach and the ontological semantic approach to lexical semantics. It is fair to say that lexical semantics, the study of word meaning and of its representation in the lexicon, experienced a powerful resurgence within the last decade.2 Traditionally, linguistic semantics has concerned itself with two areas, word meaning and sentence meaning. After the ascendancy of logic in the 1930s and especially after “the Chomskian revolution” of the late 1950s-early 1960s, most semanticists, including those working in the transformational and post-transformational paradigm, focused almost exclusively on sentence meaning. The modern movement known as lexical semantics restores word meaning as the center of interest, without, however, reviving the theoretical perspective, i.e., reasoning about the nature of meaning, or methods, e.g., lexical fields, componential analysis of word meaning, etc., of their predecessors.3 The major theoretical thrust in lexical semantics has shifted toward generativity, possibly, with a view of extending the generative approach to grammar into the lexicon, that is, toward seeking ways of grounding lexical semantics in generative syntax. There is a thirty-year-old tradition in transformational syntax of “encroaching” into the domain of semantics, often without recognizing 1 Also of the Interdepartmental Program in Linguistics and its Natural Language Processing Laboratory, Purdue University, West Lafayette, IN 47907-1356. 2 See, for instance, Frawley and Steiner (1985), Benson et al. (1986), Cruse (1986), Computational Linguistics (1987), Boguraev and Briscoe (1989), B. Levin and Pinker (1991), Boguraev (1991), Zernik (1991), Pustejovsky (1991, 1993, 1995), Pustejovsky and Bergler (1992), Briscoe et al. (1993), Dorr (1993), Saint-Dizier and Viegas (1995), Dorr and Klavans (1994/1995, 1995), Viegas et al. (1996a). 3 See, for instance, Dolgopol’skiy (1962), Garvin et al. (1967), Goodenough (1956), Greenberg (1949), Kroeber (1952), Lounsbury (1956), Masterman (1959), Vinogradova and Luria (1961), Shaykevich (1963), Trier (1931), Weisgerber (1951), Zholkovsky et al. (1961).