Semantic Fields

[...] Theories of semantic fields assume that the vocabulary of a language is structured, just as the grammar and phonology of a language are structured —the words of a language can be classified into sets which are related to conceptual fields and divide up the semantic spaceor the semantic domain in certain ways. The term ‘theory’ is loosely used; perhaps “field approach” might be preferable, since most field studies are not complete enough or formalized enough to be considered coherent, unified theories.

The development of field theories stems from the work of German linguists of a half century ago and that of American anthropologists. Both groups were influenced by Humboldt (via Boas and Sapir in the latter case), and Saussure’s notion of ‘association’ may also have been influential. Among the German linguists Trier was most important and influential. Trier distinguished between lexical and conceptual fields, whereby the lexical field divides the conceptual field into parts, like a mosaic. A word acquires its meaning by its opposition to its neighboring words in the pattern.

In applying his semantic theory, Trier investigated the field of intellect at various stages ofmedieval German, comparing different historical periods and the changes that took place.

Trier has been criticized for making a number of assumptions which may not be well founded —at least they are highly controversial. If, however, these assumptions are reformulated as empirical questions or problems to be investigated, they turn out to be very interesting problems indeed. One important question is whether in fact the lexical items in any given field divide up the conceptual space like a mosaic or whether there are gaps and overlaps. I think that the examples that are presented in following chapters will show that there very definitely are gaps and overlaps, but looking closely at the semantic field, it is possible to describe the gaps and overlaps.

A more useful model for the analysis of semantic fields is that of focal points, as suggested by the work of Berlin and Kay. In their study of color terms Berlin and Kay found that speakers disagree among themselves as to where to draw the line between colors, e. g., red and orange; moreover, the judgments of a single speaker differ at various times. However, there is high agreement and great reliability when subjects are to find focal points for colors —e. g., the most typical red or the best example of yellow. This model allows for fuzzy borders among lexical items, and thus accounts better for the behavioral evidence. The study by Berlin and Kay also shows that there are some parts of a color spectrum that are not happily covered by any term or at least by any basic term.

Trier believed that linguistic fields are not isolated, but rather that they “join together to form in turn fields of higher order, until finally the entire vocabulary is included”. There certainly are connections between fields; for example, one could establish a field of occupations, one of recreations, one of learning, and then group them into another giant field of human activities. Some fields are mutually exclusive, as for instance the fields of animals and artifacts. If something belongs to the class of animals, it is not an artifact and vice versa, notwithstanding apparent counter-examples like a toy dog which is an artifact in the shape of a dog and borderline cases like stuffed moose heads. Whether or not a progressive synthesis of small fields into larger ones is semantically enlightening is an open question.

Trier distinguished between lexical andconceptual fields, though he did not always separate the two in his writing, but many scholars have pointed out that conceptual fields cannot be defined independently of language. It is not at all clear just what the relationship between language and conceptualization is. The strong version of the Whorfian hypothesis, which claims that speakers are at the mercy of their language and can never see the world in any other way, is untenable, as is the opposite view that language has no effect on our conceptualization. The study of linguistic fields should prove to be a rich source of hypotheses about human conceptualization, and perhaps some day linguists will generally agree that the correct or at least best semantic analysis is one that describes a speaker’s conceptual structure. But at this stage, the relationship between language and thought must be considered an open one.

Another criticism that can be made of Trier’s field theory is that it assumes that lexical fields are closed, well-defined sets. Investigations of various semantic fields show that this objection is well founded. On the whole, the smaller and more specific the field, the more agreement there is among speakers on what words belong to the set. There is also agreement on the basic items in a field, whereas the disagreement is greater with respect to peripheral items —in fact many speakers may not be familiarwith the peripheral items. For example, in the field of cooking words we have bake, boil, fry, etc., but scald, carmelize, render and clarify are peripheral —according to the Berlin-Kay criteria. However, an adequate semantic analysis should reveal the difference between basic and peripheral terms and attempt to characterize theindeterminacy of the domain. Since thefirst step in analyzing a semantic field involves selecting an inventory of items in that field, the methodological problem of obtaining a valid and reliable set of words deserves attention.

Finally, Trier has been challenged for assuming that lexical fields can be organized into neat rigid patterns based on oppositions and differences, whereas the associations of words are “looser, more complex, and more unpredictable than most field theorists are prepared to admit”.

There is evidence for the view that semantic structures can be looked at in a variety of ways. [...] Within linguistics the field theory has been developed most thoroughly and explicitly by John Lyons, although other semanticists refer to or presuppose many notions of semantic fields. Lyons gives considerable importance to the notion of context, although ‘context’ turns out to be a fairly fluid concept, similar to Firth’s notion of “restricted code”. The main point of Lyons’ theory is that a lexicologist need not analyze an isolated word to account for all possible meanings and uses in all possible linguistic and nonlinguistic contexts; rather the analysis is restricted to the meaning and use of a word (and related ones) within narrow domains, as for example, a single text, the writings of a single author, a single genre, or a genre at a particular time; or the domain can be restricted by subject matter, such as looking at a set of terms as they pattern in a limited “universe of discourse”. For example, in baseball terminology, fly and run have certain meanings, and the lexicologist need not concern himself with their meanings in other contexts. The analysis of a lexical field includes the relationship of words that contrast paradigmatically (all belonging to one part, of speech), and those of other parts of speech that are related morphologically and semantically, for example the relationship between fly and wing or tutor, tutorial, and tuition.

Lyons defines the meaning of a term as a function of its relationships to the other terms in the lexical field, and these relationships (synonymy, antonymy, class inclusion, etc.) are primitives in his theory. [...]