There's millions of them: hyperbole in everyday conversation

This paper examines the occurrence of hyperbole in a five-million-word corpus of everyday English conversation (the CANCODE corpus). Hyperbole (also referred to as exaggeration or overstatement) has been studied in rhetoric and in literary contexts, but only relatively recently in banal, everyday contexts. It is often associated with irony, but the present paper also examines it in the broader context of exaggerated assertions for a variety of types of interpersonal meaning. The paper emphasises the interactive nature of hyperbole: listener reaction is crucial to its interpretation and the success of hyperbole depends on the listener entering a pact of acceptance of extreme formulations, the creation of impossible worlds, and/ or apparent counterfactuality. Corpus extracts from concordances generated for key lexical items within core semantic fields such as time and number are used to illustrate hyperbolic expressions in context, and the hyperbolic instances of the key items are identified according to a list of criteria. Figures are given for the degree of hyperbole-proneness of items, and the syntactic environment is also addressed, along with the clustering of particular kinds of signals. The paper concludes that an interactive approach to hyperbole is indispensable for its proper understanding, and that the use of large corpora offers new insights with theoretical implications for the study of tropes. (C) 2003 Elsevier B.V. All rights reserved.