Delahunty and Garvey
Some models and explanations of word meaning
Since published dictionaries do not offer a very useful model of our lexical
competence, linguists have struggled to present more plausible ones. Besides
having to account for the observations noted above, they must also explain the
fact that, while the human brain is finite, an individual’s vocabulary may be
very large. Estimates for an educated person’s vocabulary run anywhere from
50,000 to 250,000 words. The largest unabridged dictionaries of English con-
tain well over half a million entries. Clearly, however, no two individual speak-
ers of a language have exactly the same vocabulary. If this is so, how can we
hope to describe the vastness and variability of lexical competence? A general
solution is to describe not the vocabulary of a single individual or the entire
word-hoard of English, but instead to envisage the general properties according
to which the vocabulary of any individual—or of any language—can be con-
structed. There are two basic models of lexical structure, the network model
and the componential model.
The network model
The network model (N-model) posits that semantic competence is to be ex-
plained on the assumption that words have certain primitive semantic rela-
tions with each other. In other words, our semantic competence does not
consist of knowing definitions at all, but rather of knowing how words relate
to each other. You may recall from your literary theory classes that this is close
to the Saussurean/structuralist approach. The primitive relations most com-
monly explored in the N-model are the ones we’ve been discussing and are
listed and exemplified again in Table 1.
relationship characteristics examples
Synonymy extensive overlap large/big
in meaning chase/pursue
Antonymy oppositeness of meaning large/small
along related dimensions strong/weak
Hyponymy meaning inclusion rose/flower
Partonymy/Meronymy part-whole relationship keyboard/laptop
Metonymy co-elements in a situation writer/book
Metaphor similarity foot of person/
foot of bed
table 1. lexical relations recognized in the network-model
Although there are many other lexical relations, these are the most fre-