Pattern Grammar: A Corpus-Driven Approach to the Lexical Grammar of English (review)

Abstract

Reviewed by: Pattern grammar: A corpus-driven approach to the lexical grammar of English by Susan Hunston, Gill Francis Dirk Noël Pattern grammar: A corpus-driven approach to the lexical grammar of English. By Susan Hunston and Gill Francis (Studies in corpus linguistics 4.) Amsterdam & Philadelphia: John Benjamins, 2000. Pp. xiv, 289. Though lack of empirical evidence for the autonomy of syntax has certainly contributed to the decrease in popularity of the theory propagating it, the currently prevalent view that syntax is semantically motivated was not arrived at inductively: the replacement of the old dogma by the new one did not make linguistics less of a deductive science than it has been. Present-day linguistics is rife with assertions about the semantic import of syntactic choices whose only claim to empiricism is that they are spiced with a few well-chosen examples, often made-up rather than attested, which means that such statements are only checked by the intuitions of their audience, an infamously unreliable sort of evidence. This book by two members of the University of Birmingham’s Corpus Linguistics Group (formerly led by John Sinclair, one of the pioneers of British corpus linguistics) does not question that syntax and semantics are related, but the results of the inductive approach it advances force one to adopt a restrained view of their relationship. The book’s central concept is the pattern, defined as ‘a phraseology frequently associated with (a sense of) a word’ (3). In practice, the patterns discussed are for the most part complementation patterns: They describe what follows a particular verb, noun or adjective. Ch. 1 offers ‘A short history of patterns’, starting with the research of A. S. Hornby that inspired the Oxford advanced learner’s dictionary (1st edn, Oxford: Oxford University Press, 1948), moving on to some relevant research on ‘lexical phrases’ in lexicography, language teaching, and psycholinguistics, and from there to John Sinclair’s thoughts on the association of meaning and pattern (laid down in Corpus concordance collocation, Oxford: Oxford University Press, 1991) and how these were taken up and developed by Gill Francis. Ch. 2 shows how concordance lines can be used to identify patterns in a large electronically stored corpus of written and spoken texts, explains how patterns are represented in the book (using only the simplest and most superficial word-class labels like ‘v’ for a verb group and ‘n’ for a noun group instead of functional labels such as object, complement, and adjunct) and lists the kinds of pattern there are and the forms they can take. Ch. 3 deals with ‘Problems in identifying patterns’, arguing that frequent co-occurrences of words do not necessarily indicate the presence of a pattern. Chs. 4 and 5 are no doubt the most relevant to current linguistic theory because they treat the association of patterns and meaning. This association reveals itself in two ways: (1) Different senses of words tend to be distinguished by different patterns; (2) Particular patterns tend to be associated with lexical items that have particular meanings, though it is not the case that a single pattern occurs with words of a single meaning, and not all words sharing a meaning necessarily also share a pattern. Hunston and Francis warn, therefore, that the view that a word has a particular pattern because it has a particular meaning is [End Page 345] as yet insufficiently substantiated. Ch. 5 contains a comparison with the work of Beth Levin. Chs. 6–8 cast doubt on some entrenched orthodoxies in linguistics. In Ch. 6, H & F motivate their use of formal rather than structural labels by arguing that the latter only very rarely contribute anything useful to the analysis and that only the former approach can consistently deal with all verbs (and other words) in all patterns. Ch. 7 argues that word classes are best defined as pattern sets, instead of notionally or morphologically. Ch. 8 looks at patterns from a text-linguistic perspective, offering linearity as an alternative to constituency. Ch. 9, finally, considers some of the implications of an inductive pattern approach for theories of grammar and language teaching. Such an approach indeed not only challenges much of the accepted wisdom in...