In attempting to deal with referential meanings we are constantly troubled by problems of figurative extensions of meaning, for many words possess domains with marginal protuberances which do not seem to fit into the regular patterns. For example, it is not too difficult for us to treat the central domain of dog, when we are dealing merely with the different species of canis familiaris, but we are often rather hopelessly lost when we branch out into figurative extensions, e.g. (1) a despicable fellow (he's a dirty dog), (2) constellations, the Great Dog and the Little Dog, "situated near Orion, (3) mechanical devices for gripping or holding something, (4) an andiron, (5) pretension ( he put on the dog), and (6)  ruin (he went to the dogs.) Nevertheless, these various extensions of meaning are part and parcel of the semantic structure of dog.

Figurative extensions of meaning arise primarily from the process of selecting one or more components of the meaning of a particular term (e.g. physical appearance, psychological disposition, spatial relationships as in part-to-the-whole, or functional similarity) and extending them to cover some object which has not been within the domain of such a word. If an object comes to be included permanently within the domain of a particular word, there is no longer an active figurative extension (i.e. a metaphor), but simply an increase in the area of meaning of the term in question.

The interpretation of figurative meanings must carefully consider these componential features, for whatever logical validity these figurative extensions may have is based entirely upon the shared componential features and not upon any total identity. For example, a mighty fortress is our God does not mean that God is literally a fortress, but that certain features regarded as characteristic of a fortress, e.g. strength, protection, safety, unassailableness, are also the qualities of God. In general such figures of speech are built upon some feature recognized by the people of a particular speech community as being dominant, e.g. wolf (prowling rapaciousness), fox (sneakiness, cleverness), pig (gluttony), skunk (foul repulsiveness). At times entirely different symbols may have roughly similar metaphorical value, for they share certain qualities which, when extended in figurative language, become approximately the same, e.g. baloney, applesauce, and nuts, all of which can be used to mean rejection of something as worthless.

Not all societies or speech communities make the same extensions. We, for example, do not employ antelope in metaphors, though this word is widely used in metaphors in Africa. Moreover, different languages may use the same term with quite different metaphorical significance. In American English the term coon (a shortened form of raccoon) is a contemptuous figurative name for Negroes, and on the basis of this association produces a number of further figurative extensions. The equivalent of raccoon in some American Indian languages of the Middle West is used with metaphorical extensions which involve culturally approved dexterity and intelligence. Languages also differ greatly in their styles of figurative expressions. For example, most animal metaphors in English. e.g. fox, rat, bear, walrus, bull, and worm, refer to presumed psychological characteristics. In Zuñi, however, the pre-Columbian metaphors employing animal words referred to physical characteristics, e.g. an individual's hands are "rough like the feet of a turkey"; someone is "as bald or downy-headed as an eaglet"; his eyes "protrude like a rat's"; he has "bony legs like a bird" (Stanley Newman,1954).

Languages may also differ in the extent to which they employ metaphorical extensions, especially in certain areas of the vocabulary. In English, for example, wẹ have a number of metaphors based on animals (e.g. fox, rat, ass, goat, monkey), insects (e.g. louse, bug, fly), flowers (e.g. pansy, lily, rose), vegetables (e.g. tomato, beans, onions), and fruits (e.g. banana, apple). However, English has nothing like the number of metaphors common in Brazilian Portuguese, in which almost all animals, fruits, and vegetables have certain metaphorical extensions of meaning, many with vulgar connotations.

Languages differ also in the readiness with which they admit new figures of speech. For example, Tarascan , a language of Mexico, which has a number of metaphors, does not readily admit new ones. In contrast, Cuna, a language of Panamа, not only has many metaphors but admits new ones readily, apparently with considerable cultural approval, and even avidity.

An essential part of the communicative power of figures of speech is derived from the central meaning of the word, which still continues as an active force. Once the central meaning-which provides the basis for extension of some componential quality-is lost, the figure also loses its force; for the strength of the figure lies in the relationship established between the central, or core, meaning and the extension.

责任编辑：admin