- Antonymy: A corpus-based perspective by Steven Jones
A revision of Jones’s 1999 Liverpool doctoral thesis, this book provides a fresh and interesting perspective on antonymy, applying a corpus methodology in order to make discoveries about the uses of antonyms in discourse. In contrast to traditional approaches to paradigmatic lexical relations which posit logical definitions or make claims about antonym substitutability, the corpus methodology allows J to focus on the roles that antonymy plays in texts. J classifies these textual or rhetorical functions of antonymy into two major and six minor classes. One major class is ‘co-ordinated antonymy’, in which antonyms cooccur on either side of a lexical (or understood) conjunction, as in both listed and unlisted properties or He would neither confirm nor deny the rumor. These account for about 38% of antonym cooccurrences in J’s newspaper corpus. The other major class is ‘ancillary antonymy’, in which two pairs of items are contrasted in the same sentence, one of them an established set of antonyms and the other a less-established contrasting pair, as in It is meeting public need, not private greed. In this example, the established antonymy of public/private supports the perception that need/greed form an antonymous pair as well. In J’s corpus, up to 40% of sentences with cooccurring antonyms have two contrasting pairs in this way.
For these and the minor classes, J identifies a number of construction types (or ‘frameworks’ in his terms) in which antonyms cooccur. Because these are productive, in that different word pairs can be inserted, this approach provides a means to explain how antonym relations develop diachronically. Since the constructions themselves can be said to carry a contrastive function, antonymy is established through repeated use of a pair in a variety of such constructions.
In addition to providing a taxonomy of rhetorical functions for antonyms and of the lexicosyntactic constructions in which they occur, J demonstrates the endemicity of antonyms (suggesting that as many as one in 50 sentences contains cooccurring antonyms) and calls into question traditional notions of antonymy as a paradigmatic (rather than syntagmatic) relation. Because the corpus used (280 million words from the British Independent newspaper) is severely limited in the range of discourse types and authors involved, the findings raise many questions. For example, are the ranges and proportions of antonym functions similar in other registers? In speech? In other languages and cultures? Can changes in antonymic status be observed in diachronic corpora, and [End Page 840] do they indicate that certain functions or frameworks are most effective in establishing a pair’s antonymy? Which comes first (in acquisition and semantic change), the frameworks or the antonyms? Since this book provides the vocabulary and the methodology for asking and answering these questions, it is both an important contribution to lexicology and a demonstration of the usefulness of well-designed corpus research.