A helpful word from Louw-Nida, whose own lexicon (based on semantic domains) represents an important step forward in Greek lexicography:
Rather than grouping glosses into seemingly related sets, though often in a somewhat haphazard manner, it is much more satisfactory to define meanings in terms of those distinctive features which may be said to determine the range of usage of a lexeme (namely, a word or an idiom). Most people, however, are so entranced with the usefulness of words that they assume that in some way or other the words used in speaking about certain phenomena must provide a clue as to their true nature. But words are not labels for the contents of concepts. They are simply signs or symbols by which entities, activities, characteristics, and relations are represented in the process of communication. And this is possible only by virtue of the fact that verbal signs are all part of a system of signs. . .it is this system which makes meaning possible, since the signs of any code are only defined by other signs
Words are vocalized signs whose meanings are interdependent and indeterminate since they fluidly represent a referent, instead of labeling it in a fixed sense. There is no direct relation between a symbol and its referent. Rather, there is a structural relation between a symbol and the sense associated with it. Moreover, in order to define signs, we must use signs, thus complicating further the problem of locating and delimiting the semantic field of any given sign.
Words have no inherent ‘meaning’ (though some may denote an actual object). Words have relatively stable semantic fields based on the scope of their various uses. In fact, it is ‘meaning’—a feature determined by context (syntagmatic, literary, situational)—with which we normally associate words (think of phenomena like deixis and implicature). As Nida writes, “Meaning consists of that particular structured bundle of cognitive features, associated with the lexical unit, which make possible the designation of all the denotata by the lexical unit in question.” An utterance, then, is coherent when its words faithfully reproduce the mental representation intended by its author. This is the goal of words. Louw-Nida’s lexicon often errs, but it does so with an emphasis on context, which, to borrow from Barth, is “an error in the right direction.”
When assembling a lexicon, the lexicographer must operate from a number of fundamental assumptions about language. I want to include Louw-Nida’s here, since their assumptions are especially helpful in understanding the difficulties involved in lexical semantics:
1. Since languages have a limited number of verbal signs (words and idioms) with which to represent an unlimited number of entities, activities, characteristics, and relations, multiple meanings for many lexemes are inevitable.
2. Languages consist of open systems, since lexemes may be added or lost and the range of their meanings may expand or contract. Without this feature of language there would be no figurative use of language and no poetry.
3. The boundaries of meanings are indeterminate in the sense that they can be vague and have fuzzy edges. For example, it is impossible to tell how thick a string has to be before it should be called a cord, or how thick a cord must be before it should be called a rope.
4. As in the case of all systems, languages are incomplete and have anomalous features, which can perhaps be best described as involving “parallax,” a kind of systematic distortion such as occurs in all maps of the earth and in all photographs. Since languages reflect culture and culture is often unsystematic, languages also represent cultural distortions, e.g. sea lions are not lions, ringtail cats are not cats, and the evening star is not a star but a planet.
5. The meanings of verbal signs are determined by other verbal signs, and this means that ultimately there is no such thing as an absolute definition of the meaning of any sign, although there may be practical definitions.
6. A combination of verbal signs is never the same as the meanings of the individual lexemes, since in addition to the lexical meanings there are also meaningful syntactic and rhetorical relations.
7. The meaning of a sign is the minimum of what that sign contributes to the context. This represents the principle of entropy (often spoken of as the second law of thermodynamics) in which the significance of the context is maximized and the role of the individual element is minimized.
8. The meanings of lexemes are not equivalent to reality, but only represent the manner in which the speakers of a language perceive reality—an important implication of the sociology of knowledge.
9. Modern taxonomic classifications are not based on a concept of rigid “pigeon-hole” distinctions based on “necessary and sufficient features’* but on the concept of family resemblances, bundles, clusters, continuums, and multidimensional relations. As a result, definitions must often express uncertainties by such expressions as perhaps, probably, sometimes, in general, often, and usually.
These basic assumptions about language and methods of semantic analysis have several practical Implications which need to be considered before attempting to analyze lexical meanings. The nature of language inevitably influences the manner in which a person should go about the analysis of lexical meaning.
Nida, Eugene A., and J. P. Louw. Lexical Semantics of the Greek New Testament (Atlanta: Scholars Press, 1992). 17.