Corpus Linguistics at Work (review)

The title of this book would seem to indicate that it is about applications of corpus methodology, and since it starts out (in Ch. 2) with a discussion of the value of the corpus methodology in language teaching and learning, one is not quickly disabused of this notion. However, its main thrust is to argue for a corpus-driven linguistics (CDL) versus a corpus-based approach. While such an approach is not new (indeed, one could say it goes back to the Firthian contextual theory of meaning, and has been argued for particularly by John Sinclair), it is only in recent decades that both the technology and the available corpora have allowed its serious practice. The time is ripe for a cogent argument for CDL and some guidelines for practicing it. Unfortunately, the organization and multiple purposes of this book often keep it from reaching this ideal in that it veers among definitions, arguments, and exemplifications, but not in an order that serves to guide the reader through the material.

After an introduction (Ch. 1), the book begins with an application of corpus methodology (Ch. 2), then defines issues in the use of corpus evidence such as authenticity and representativeness (Ch. 3). It then compares the corpus-based approach and CDL (Chs. 4 and 5), discussing ways in which the former limits the empirical potential of corpus data—for example, by analyzing data in terms of predetermined theoretical categories—and demonstrating this limitation by describing and replicating Sinclair’s argument that corpus evidence argues against treating inflected word forms as equivalent to instantiations of lemmata. That is, differently inflected forms (e.g. faced and facing) occur in very different lexical and grammatical contexts. After this, the issues of what can constitute a meaningful unit in language are discussed (Ch. 6), with definition and exemplification of ways in which words become less semantically independent in context (semantic prosody, delexicalization) and less evaluatively neutral. In Ch. 7, the author turns to the usefulness of corpora in determining translational equivalence (for comparative/con trastive linguistic inquiry and translation). Next she returns to theoretical issues, describing the Firthian approach to meaning that has underpinned her approach in the previous chapters (Ch. 8) and then giving a history of other theories of meaning (Ch. 9). The conclusion (Ch. 10) includes a useful position statement for CDL, then looks back and considers the issues raised in the book, but it does so in a more logical order (theoretical background and history, motivation and explication of CDL, and demonstration of its applications).

The book contains a great deal of thought-provoking discussion and data and so can be recommended to those who are interested in practicing or critiquing CDL. I do, however, recommend that readers start with Ch. 10 and then follow the alternative chapter order for the book proposed on page 180 (possibly omitting Ch. 9). While the book draws on a great number of other works, it cannot be recommended as a research starting point because of the huge number of errors of omission, mismatch, or underspecification in the citations and bibliography. (I counted nearly 60 before I stopped counting and am coming to think of such inattention to editing as the hallmark of this publisher.)