Representation of a Polysemous Word in Models of Mental Lexicon

The article reviews some of the most prominent theories of mental lexicon in regard to the problem of polysemy representation. The focus of the studies in the field of mental lexicon is on the search for an adequate way to represent, store and retrieve from the memory information about the senses of a polysemous word. The advantage of cognitive models of representation over "listing" senses in the lexicon is that they view a polysemous word as a multi-faceted structure of world knowledge rather than as a set of isolated senses. In this respect, a сognitive approach to polysemy proposes models of lexicon integrating dictionary and world knowledge. From the point of view of representation of such knowledge, the advantage of connectionist models of verbal memory is discussed, special emphasis is made on the associative verbal model. The existing studies of polysemy by the method of free associations have not investigated the semantic structure of a polysemous word as a cognitive phenomenon so far. The established method of analysis of responses to polysemous stimuli aims to differentiate the senses of the word in the associative field, since it is based on the assumption that distinction between "senses" and "meanings" is irrelevant for a language speaker. The relevance of the present study consists in the attempt to model polysemy of the word as a cognitive phenomenon. The author's aim is to propose a method of analysis of associations that allows to state the different status of homonymous lexemes and senses of a polysemous word in the lexicon. The results discussed in the article were obtained from the responses to polysemous and homonymous stimuli elicited from Russian-speaking students during a free association experiment. The analysis of the data involved semantic interpretation of the responses, their clustering and construction of the associative field. As a result, the structures of associative fields of polysemous and ho-monymous stimuli were compared. The method of analysis is aimed at the search for grounds on which senses of polysemous words are perceived as related by speakers of language. The comparison of the structures of polysemous and homonymous words revealed the peculiar features of both homonymy and polysemy. For instance, for the homonymous stimulus ("ssylka") it was found that associative responses are distinctly and more easily distributed into two groups referring to different meanings of the word (ssylka1 "link, reference" and ssylka2 "deportation"), while the same differentiation is more difficult to accomplish when it comes to the senses of the same word (ssylka1 as "link" or "reference"). The discovered difficulty is due to the responses which cannot be univocally attributed to a particular meaning (sense) of the word. These responses should be regarded as indicators of interrelatedness of the senses of the word, since they are more typical of polysemous stimuli rather than of homonymous. The analysis of the associative field of the polysemous word "set' " ("net") has also shown that the difficulty in attributing of association to one of the senses denotes the complex structure of senses typical of polysemy. Thus, the established feature of the associative field of a polysemous words presumably reflects their representation in the lexicon, different from homonyms.

Keywords

Authors

References