Progress in English computational lexicography

Abstract

: This paper is a progress report of the Collins COBUILD English Language Dictionary project. The project, initiated in 1979–1980, pioneered modern computing techniques in lexicography research. The acronym COBUILD used for the project stands for Collins the publishers, and Birmingham University International Language Database. This report discusses the corpus—a collection of 20 million words of text—and presents the historical, theoretical and methodological bases for the project. This report briefly discusses how, for example, the following issues concerning the data were tackled: meaning, grammar, definitions and explanations, and examples. In addition to discussing the main compilation process, this paper also discusses how structural forms of the language are related to meanings and uses, and how considerations of frequency affect the description. The special problems related to indication of stress and rhythm are also discussed. This paper concludes with observations on the wider implications of the COBUILD project.