Antonymy is the technical name used to describe 'opposites', pairs of words such as rich / poor, love / hate and male / female. Antonyms are a ubiquitous part of everyday language, and this book provides a detailed, comprehensive account of the phenomenon. As well as re-appraising traditional semantic theory and re-evaluating existing categories of antonymy, the book raises wider issues, including questions such as: Where do new atonyms come from? Which pairs can be regarded as 'good opposites'? Why do atonyms tend to favour a particular sequence in text? Does word class affect the function of antyonymy? These questions are addressed from a corpus-based perspective, using statistical evidence from a database of 3,000 sentences retrieved from a 280-million word corpus of newspaper text. Fresh, empirical statements are made about antonymy, and over 350 authentic examples of 'opposites' are recorded and analysed. This book demonstrates how traditional linguistic theory can be revisited, updated and challenged in the corpus age. It will be essential reading for scholars interested in antonymy and corpus linguitics.