Content area
Abstract
This dissertation studies lexically-specific (irregular) polysemy, using a case study of the English verb see as the major example. Clearly, words such as see have different meanings in different contexts, but how can we distinguish different senses from mere different uses (modulations) of the same sense? What are the semantic and paradigmatic relations among the senses? Answers to these questions were sought through a series of psycholinguistic experiments, formal analysis in terms of Frame Semantics (Fillmore 1976; Fillmore 1982; Fillmore & Atkins 1992) and other cognitive linguistic theories, and analysis of entries in monolingual and bilingual dictionaries. The results show that speakers can reliably distinguish many senses of see, that the English pattern of senses is partially shared across languages, and that frame semantics is a good way of representing the relations among senses.
First, the relation of semantics to world knowledge and categorization is discussed, and Frame Semantics, homonymy, polysemy, and monosemy, traditional tests for polysemy, and other types of linguistic evidence are defined.
Then, the semantics and syntax of see are outlined and detailed frame representations are given for 19 senses (e.g. RECOGNIZE ( saw that he left), ENSURE (see ( to it) that he leaves) and EXPERIENCE (saw combat)) and 6 purely compositional uses, e.g. TOUR and HALLUCINATE, including inheritance (complete or partial) among senses and from more general frames. Representative collocations with see are discussed, along with other cognitive linguistic representations, including mental spaces (Fauconnier 1985 [1994]).
Next three psycholinguistic experiments are described, involving (1) free sorting of examples of uses of see, and (2) untimed, and (3) timed classifying of examples into a priori categories. Results suggest that speakers can reliably access finely differentiated senses like those proposed above.
The sense divisions for see in entries from several English dictionaries are shown to be problematic. Entries from bilingual dictionaries between English and Spanish, Chinese, and Japanese demonstrate that there is partial overlap between the senses and sense inheritance of English and those of other languages. The similarity of English to Spanish is greater than to the non-Indoeuropean languages.
Finally, conclusions are drawn from all the approaches and future research is outlined. Includes index and experimental stimuli.
