DescriptionPublished as Coyote Papers: Working Papers in Linguistics from A-Z, Exploring Language: Linguistic Heresies from the Desert
AbstractThe theoretical activities and insights of the last two decades in linguistics have not spilled over into etymology and the study of semantic change, even though there has been much important work in both historical linguistics and in semantics. One reason for this neglect of semantic change is that the changes themselves seem to be sporadic. Every word has its own history. About the best we have come to hope for is a taxonomy, or classification schema, as found in Ullmann (1957), Stern (1931, 1968), or Williams (1975). These categories of semantic change summarize the tendencies or possibilities which may in fact have opposite effects, as narrowing vs. broading. Our current state of knowledge does not allow us to state interesting, falsifiable statements concerning the lexicon as a whole. In this paper we shall argue that some insights into the principles of semantic change can be found by looking, not at the whole lexicon, but at words which belong to a single semantic field. A semantic field is a set of lexemes which cover a certain conceptual domain and which bear certain specifiable relations to one another. An example of a simple semantic field would be the conceptual domain of cooking, which in English is divided up into the lexemes boil, bake, fry, roast, etc. A basic premise of semantic field theory is that to understand lexical meaning it is necessary to look at sets of semantically related words- -not simply at each word in isolation. By 'semantically related' we refer to relationships between lexical items such as synonymy, as in big and large; antonymy; such as big and small, hyponymy, as rose and flower or robin and bird; converseness, as buy and sell; incompatibility, such as cat, dog, cow, horse, pig, etc. A list of such lexical relationships and their meaning can be found in Lyons (1977) or Lehrer (1974). We will show that our understanding of semantic change can be enriched by looking at the histories of semantically related words.