Review: Gries (2009) Quantitative Corpus Linguistics with R. London and New York: Routledge

Abstract

There are many teachers and scholars who value corpus analysis, but do not yet do their own programming for teaching and research. Many have been considering making the leap from using only pre-packaged concordancing software to writing our own programs, and Gries’s Quantitative Corpus Linguistics with R may be the resource that helps many of us to make the transition. It may be difficult, however, to see the advantages of doing one’s own programming until one has conducted corpus research using only concordancers and developed a sense of their limitations. First, I will allow that concordancers are very useful when the focus of the research is on very specific, pre-determined words or phrases, or other top-down sorts of research questions. Concordancers are a good way to introduce students to the concept of corpus-informed language learning and they demonstrate the way that near-synonymous lexical items vary in their functions. These uses have been garnering increasing attention in the wider fields of Applied Linguistics and even Theoretical Linguistics, and we can thank concordancing software for making corpora more accessible and opening up the tool of corpus analysis to an increasing number of students, teachers and researchers. However, when one starts to ask deeper questions about patterns of language use, and we desire a more bottom-up approach to linguistic analysis that is not dependent on pre-determined lexical or grammatical items, the limitations of concordancers quickly become apparent. To limit our linguistic inquiry only to questions that can be investigated through pre-packaged software is akin to limiting our study of syntax to items that can be identified by the green squiggly line of the grammar checker in our word processor. We are at the mercy of the software developers in terms of the analytical features included and the price to purchase or upgrade these tools.