Corpus-Based Language Studies: An Advanced Resource Book (review)

The corpus-driven revolution in applied linguistics continues apace, and along with it the paradox that while corpora are changing the face of applied linguistics (most dictionaries, grammars, and course books now claim to be corpus based), this is occurring largely without the participation of practitioners. Only a few teachers or researchers have ever built a corpus or delved through concordance lines. Possibly in a bid to remedy this hands-off problem, the book display at the recent conference of the American Association of Applied Linguists (AAAL) in California in spring 2007 offered roughly a dozen books promising to enlist language teachers and graduate students in the revolution at a more basic level. The two I walked off with were the volume under review here and O'Keeffe, McCarthy, and Carter's (2007) From Corpus to Classroom.

It quickly became apparent that the latter volume was basically a set of conclusions about how corpus evidence can inform teaching practice, based on insights drawn from the impressive Cambridge and Nottingham CANCODE corpus of written and spoken British English, but that there would be no hands-on opportunities to construct further insights of one's own from this corpus, owing to its total inaccessibility to the broader professional community. So it was a relief to find that access and hands-on were main themes in the other volume.

McEnery, Xiao, and Tonio's Corpus-Based Language Studies is one in a new series by Routledge/Taylor & Francis dealing with themes in applied linguistics (intercultural communication, translation, grammar and context, and SLA) and bearing the common subtitle 'an advanced resource book.' Access and support for hands-on projects are main themes of the series and are arguably key to a new-approach show of force from a publisher (the new owners of Lawrence Erlbaum) bidding to enter the language education field as competition for Cambridge University Press. So how accessible and hands-on does the McEnery et al. volume get?

The series format involves presenting first concepts and procedures, then related excerpts from actual research articles, and finally fully worked example research projects that can be replicated. Topics include corpus types, corpus building, corpus tagging (adding part-of-speech, or POS, markers), corpus statistics, corpus controversies, and corpus analysis for various purposes. Research article excerpts include a thoughtful and relevant collection of true area classics by Widdowson, Biber, Stubbs, Carter, McCarthy, and others. Research projects unpacked in detail include a methodology for corpus-based lexicography; a study in the sociolinguistics of British swearing, made possible by the huge and finely subdivided British National Corpus (BNC); and a replication using the Longman Learner's Corpus of the empirical morpheme acquisition sequences studies of Krashen and others in the 1970s-to mention fewer than half of those available. No other volume that I know of has gone so far to bring corpus research into focus for practitioners, or to make at least the beginning of a serious research project actually possible.

But there are probably limits to how accessible such an involved methodology can actually be made. For one thing, many of the book's analyses involve the use of Mike Smith's text-analysis program Wordsmith, which must be purchased, to get beyond the demo limitations, for about CAD$100. There are, of course, Web concordancers, and, indeed, the swearing study mentioned above is based on the University of Zurich's Web concordancer, which apparently runs the full BNC. Getting onto this site requires a password, however, which the book does not actually mention; when I e-mailed the site administrator, I received a irritated reply complaining about the number of users this book was sending them, as well as a reluctantly given username and password, which did not work).

The biggest access problem in this area, of course, is not software (several excellent free concordancers are available for download, such as Lawrence Anthony's AntConc) but, rather, corpora, such as the Japanese EFL...