Artigo Revisado por pares

Cross-linguistic Evidence for Cognitive Foundations of Polysemy

2014; Wiley; Volume: 36; Issue: 36 Linguagem: Inglês

ISSN

1551-6709

Autores

Huichun Zhu, Barbara C. Malt,

Tópico(s)

linguistics and terminology studies

Resumo

Cross-linguistic Evidence for Cognitive Foundations of Polysemy Huichun Zhu (huichun.zhu@gmail.com) Department of Psychology, 17 Memorial Drive East Lehigh University, Bethlehem, PA 18015, USA Barbara C. Malt (barbara.malt@lehigh.edu) Department of Psychology, 17 Memorial Drive East Lehigh University, Bethlehem, PA 18015, USA The possibility of complete unpredictability is defeated by the observation that some relationships between central sense and extended senses of words tend to recur. For instance, at least within English, the same word can be used to refer to both the object and the substance (e.g., chicken; fish); both the text and the object (e.g., newspaper; book); both the act and the instrument (e.g., drill; brush); and both the actor and the act (e.g., cook; scout) (e.g., Blank, 1999; Norrick, 1981; Nunberg, 1979). More broadly, metaphor and metonymy frequently characterize the relations between the most literal sense of a word and its extended senses (Stern, 1931; Ullmann, 1962; Blank, 1999). For instance, the foot of a mountain may be metaphorically similar to a human foot, and the tongue spoken by a linguistic community is the language that emerges by means (partially) of the tongue in their mouth. Still, such observations are largely descriptive and post- hoc. They do not indicate, for any given word, what senses are likely to arise, and they provide no explanation of why those senses may come about. At a broad level, they do suggest that language users engage in metaphorical and metonymic thinking and also that they are sensitive to more specific relations such as that of an object to its substance. But that does not reveal what particular senses are likely to be created or why. Indeed, Lakoff (1987) and Langacker (1988) have suggested that although the array of senses conventionally associated with a word is not arbitrary, neither is it readily predictable. These analyses of recurring relations have generally been carried out without closer consideration of the cognitive processes that might be involved. We suggest that further consideration of the cognitive processes contributing to generation of extended senses can help provide predictive power. In particular, we propose that salient characteristics of the default referents of the core senses (the most central, literal senses) provide the basis for sense extensions. Consider the word head. Suppose that the core sense of the word refers to a certain body part of animals. Speakers of a language know many things about typical referents of this sense of the word. But those things they know are not all equally salient or important. Although people know that heads are made of substances such as flesh, bones, blood, and so on, these features are not likely to be the ones they bring to mind when thinking about heads. Instead, properties such as located at top of body and organ of decision making are more uniquely and saliently associated with heads. In turn, these salient features will be more Abstract Existing discussions of polysemy describe the relations that extended senses may have to the most central sense of a word, but they do not explain in more detail how particular senses are generated for a given word. We propose that extended senses are initially built on the salient features of referents of core senses (and further senses may be generated from those). We provide evidence for the role of salient features of core senses in generating extended senses through three studies. These studies use speakers of English and Chinese, historically unrelated languages. Keywords: polysemy; word meaning; word senses; cross- linguistic comparison Introduction Polysemous words are those that have multiple different but related meanings (e.g., foot as in “my left foot”, “the foot of the chair,” “at the foot of the mountain”). Polysemy is the rule more than the exception for words of moderate to high frequency (e.g., Berlich, Todd, Herman & Clarke, 2003). This one-to-many mapping of form to meaning has raised questions about the representation and processing of polysemous words (e.g., Bohrn, Altmann & Jacobs, 2012; Caramazza & Grober, 1976; Klein & Murphy, 2001; Simpson, 1994). But studies addressing such questions leave unanswered a logically prior question: How do polysemous senses of words arise? At one extreme, it could be that the generation of new senses from an initial word meaning is entirely unpredictable. The first extended senses that arise for any individual word might be highly influenced by the particular lexical gaps that exist and other characteristics of the language involved (such as where it falls on the synthetic- to-isolating continuum, which affects how morphology can be used to help build sets of related senses). They may also be influenced by the cultural conditions present (e.g., introduction of new products or practices) that drive innovation for the linguistic community. The result of these influences could be that the senses tend to be idiosyncratic to that word and that language. Furthermore, because extended senses may build on one another to create chains moving farther from the original (Lakoff, 1987), once the first extended sense emerges for a word, the additional senses it spawns may set the chain for that word off in directions not followed by other languages for their most closely related word.

Referência(s)