In Part 1, MacMahon describes the work of early British linguistics (e.g., the Oxford English Dictionary), and then discusses the development of phonetics, the role of Daniel Jones, his work on the phoneme, and his belief in ‘practical’ phonetics. After describing Alan Gardiner’s work, he focuses on John Rupert Firth, the dominant figure in the early 1930s to late 1950s, who taught mainly at SOAS in London. He developed ‘a contextual theory of meaning,’ based on Malinowski’s ideas, ‘prosodic phonology’ and the idea of several simultaneous systems (polysystematicity). After Firth’s death, Chomsky’s ideas about syntax became dominant, while the term ‘neo-Firthian’ was applied to Michael Halliday and Systemic Linguistics from the early 1960s.

In Part 2, McEnery and Hardie survey neo-Firthian corpus(-driven) linguistics by John McHardy Sinclair and his group, their focus on ‘collocation’ (co-occurrence patterns of words in discourse), and their interest in ‘lexicogrammar.’ For them a corpus is central: theories and analytic generalizations and categories must emerge from corpus data. They impacted lexicography by setting up COBUILD to provide data, ideas, and analyses. Also important is the ‘Idiom Principle’: i.e., in language there are a large number of semi-preconstructed phrases, which has led to two different neo-Firthian theories of language.

Every reputable new English dictionary, or new edition, published since 1987 has made use of a large collection of text, or ‘corpus’, for evidence of a word’s usage. In this chapter, one of the most eminent names in lexicography, Patrick Hanks, takes the reader on a journey to discover more about different kinds of dictionaries and corpora, and basic principles of corpus linguistics and lexicography. He outlines how dictionaries have made use of corpus evidence in the past, and proposes how they might make better use of them in the future.

This chapter reviews the transformative effects of technology on dictionary-making, focusing on four main areas: the use of databases for storing and organising dictionary text; the creation and exploitation of corpora for use as the dictionary’s evidence base; the enhancement of the value and usability of corpus data through the application of software tools developed in the NLP (natural language processing) community; and the migration of dictionaries from print to online media. During the last half-century, activity in all these areas has brought fundamental changes to the way dictionaries are created and made available to their users. We trace the development of corpus-based lexicography in English, from the early work of John Sinclair and his colleagues in the 1980s to the present day. Lexicographers working in English and other widely used languages now have access to resources which would scarcely have been imaginable thirty years ago: very large corpora (measured in tens of billions of words) and sophisticated corpus-querying tools are routinely available. The move from print to digital publication is a more recent development, but no less significant. The far-reaching implications of these changes – for dictionary-makers and dictionary-users alike – are explored at every stage.