- Polysemy: Theoretical and computational approaches ed. by Yael Ravin, Claudia Leacock
This volume distinguishes itself by its very catholic table of contents, including a wide variety of theoretical views as well as several computational approaches to the problem of disambiguation. [End Page 207] Ravin and Leacock’s ‘Polysemy: An overview’ (1–29) gives a concise history of the progress and problems in accounting for polysemy in four approaches: classical models of definition, prototype approaches, relational models, and computational approaches. In doing so, they highlight the contributions made by the chapters in this volume to each of these traditions and thereby give a semblance of coherence to this wide range of work. It’s interesting to note, however, that the word pragmatic(s) does not appear here. Polysemy is considered a semantic problem, such that ‘The challenge for semantic theories lies in formulating mechanisms to trigger the relevant meaning components in the appropriate contexts’ (22). Given the wide range of applications of the word semantics in the diverse approaches presented here, that these triggering mechanisms should be assumed to be a semantic problem is legitimate only in the widest understanding of the term semantic: ‘having to do with meaning’.
In the conclusion of his ‘Aspects of the microstructure of word meanings’ (30–51), Alan Cruse helpfully names this context-sensitive approach to meaning ‘soft semantics’. Cruse’s chapter demonstrates the range of variation among polysemes using a variety of ambiguity and unity tests and proposes three sources of discontinuity in meaning: subsenses, facets, and ways-of-seeing. This taxonomy of different meaning types is presented against the backdrop of a view which holds that meaning is ‘infinitely variable and context sensitive’, but that ‘there are nonetheless regions of higher semantic “density”. . . forming as it were more or less well-defined “lumps” of meaning with greater or lesser stability under contextual change’ (30). Subsenses are diagnosed if two referents of the same use of the word must be of the same type and if the meanings have independent truth conditions and different sense relations. On these criteria, Cruse argues that knife has subsenses corresponding to the various types of knife (scalpel, cutlery, penknife). Unlike subsenses, facets cannot be so clearly separated. Cruse uses the example of book, whose tome and text meanings are clearly different but can be evoked by a single use of the word, as in an interesting and well-produced book. Ways-of-seeing (WOS) correspond loosely to Pustejovsky’s qualia categories and are less distinct than subsenses or facets. Cruse acknowledges that the implications of his views are not entirely worked out, and there is a fair amount of vagueness in this chapter as to the nature of the semantic muck in which lumps and nodules form. Nevertheless, his taxonomy is a useful one, and he concludes that word meanings cannot be isolated outside of context because they are ‘irreducibly complex’ and that prototypes are likely insufficient to capture the nuances of semantic unity and discontinuity.
Christiane Fellbaum’s chapter on ‘Autotroponymy’ (52–67) looks at verbs with superordinate and subordinate senses, in which the subordinate sense conflates some argument or modifier. While these types are not regular, Fellbaum finds some generalizations among them. James Pustejovsky’s ‘Lexical shadowing and argument closure’ (68–90) focuses on a much more specific area of verbal polysemy, accounting for two types of optional argument expression, risk-type verbs and cognate object verbs like dance. The chapter depends on previous familiarity with and acceptance of Pustejovsky’s generative lexicon framework.
Charles Fillmore and B. T. S. Atkins give us ‘Describing polysemy: The case of crawl’ (91–110), in which they present a semantic network of crawl’s meanings (derived from corpus data) and compare crawl to the French ramper. Their methods have obvious applications for lexicography, and they suggest reasons why semantically similar words do not follow the same polysemous patterns crosslinguistically.
The next two chapters bring into question the boundary between the syntactic and the...