This dissertation investigates what makes two words antonyms. Previous research has not adequately explained why some words seem to contrast in meaning but are still not considered antonyms (e.g. large and little) nor can it explain why some words have two antonyms (e.g., happy/sad and happy/unhappy). An explanation is given here using the notion of "semantic range" (a description of a word's typical collocation patterns); antonyms are shown to be words which have a great deal of semantic range in common.
In the first chapter, previous characterizations of antonymy are reviewed, and on the basis of their limitations, a new characterization of antonymy in terms of shared semantic range is proposed. This proposal is developed and argued for in the following three chapters, each of which is an in-depth case study. The second chapter presents the adjectives big, little, large, and small, focusing on the question of why little and large are not antonyms even though they name opposite ends of the semantic dimension of size and even though the pairs big/little, large/small and big/small are all antonyms. The third chapter examines several adjectives related to the concepts of wetness and dryness--wet, damp, moist, dank, humid, dry, arid and parched--and explains why of these, only wet and dry are antonyms. The fourth chapter explores the adjective happy and its two antonyms, unhappy and sad.
In each case study, the semantic range of each adjective is characterized in terms of the kinds of nouns which it typically modifies, and then the semantic ranges of the adjectives under consideration are compared. Adjective-noun co-occurrence patterns from a large corpus are used as the main source of data. When the semantic ranges are compared, antonyms are found to be adjectives which have a high degree of overlap in semantic range; for example, big/little, large/small, and big/small are found to share a great deal of semantic range but large and little have almost no shared semantic range and thus are not antonyms.
Many people have helped me directly or indirectly with this dissertation, but there are a few without whom I could not have finished it at all. First of all, I would like to thank the committee members who guided me through this dissertation: Beth Levin, Ossy Werner and Gil Krulee. I am especially grateful to Beth for helping me work out the problems with doing the dissertation over such a long distance (from Tokyo). I am also grateful to John Wickberg, who talked with me about my ideas at the early stages and then helped me by writing the computer program that was used to gather the data from the New York Times corpus, data which were used in much of my analysis. I'd also like to thank Gail McKoon for looking up additional examples in the corpus when I needed them.
Throughout the time I've been working on this dissertation, I've had a big "cheering team" to encourage me and to help me find time to write. I'd especially like to thank my husband, Jerome Young; my parents, Ardis and Terry Muehleisen (who kept asking "When will you be finished?"); fellow Northwestern students Grace Song and Laurel Smith Stvan (who also helped with proofreading); NU graduates Mutsumi Imai and Bill Snyder (who showed me it could be done); Jeff Stvan; my brother and sister-in-law, Ralph Muehleisen and Sally Laurent-Muehleisen; and my colleagues at Waseda University, Takashi Ida, Mitsuru Mizuno, and James Fegan.