Corpus Linguistics: Method, Theory and Practice

Citation

McEnery, T., & Hardie, A. (2012). Corpus Linguistics: Method, Theory and Practice. Cambridge University Press.

Chapter Summary

The book has nine chapters, a glossary, notes, references and an index, in about 312 pages, and belongs to the Cambridge Textbooks in Linguistics series. The preface states its purpose: to explain how corpus linguistics has developed and is used as a methodology, the major theoretical issues corpus linguists face, and the problems researchers face when using corpora in practice, within linguistics and across disciplines. The authors say they do not give step-by-step instructions for investigating a corpus and do not go into the details of concordances, collocations, keywords and other standard tool outputs, because other books cover that ground. Chapters 1 to 3 cover method and practicalities, Chapters 4 to 6 the history and main traditions of the field, and Chapters 7 to 9 its relationship to other branches of linguistics and its future.

Chapter 1: What Is Corpus Linguistics?

The opening chapter defines the field as a set of procedures for studying language, while warning that corpus linguistics is a heterogeneous field, and not a monolithic set of agreed methods, and that at least one major school does not accept the description of it as a methodology. A corpus is characterised as a set of machine-readable texts, usually too large to analyse by hand, taken as an appropriate basis for particular research questions. The chapter argues that corpus linguistics may reshape linguistic theory itself, not only supply data for it.

Chapter 2: Accessing and Analysing Corpus Data

The chapter begins with the history of data in linguistics: observed data before the mid-twentieth century, the rise of introspection under Chomsky's influence, and the rehabilitation of corpus data after 1980. It then moves to practical questions about selecting, annotating and exploiting corpus data, and includes a section on the early work of Roberto Busa, to which Chapter 4 refers back.

Chapter 3: The Web, Laws and Ethics

The chapter continues the survey of practical issues with legal considerations in building corpora, taking web-based compilation as its main case, and with the ethical issues in constructing, distributing and using corpora. The authors note that they write from a Western standpoint and do not attempt a global survey of laws and ethics.

Chapter 4: English Corpus Linguistics

This chapter reviews the tradition of work on English from the 1960s onward, which the authors call the crucible of the approach described in the book. It also acknowledges early corpus work on other languages, such as Busa's and Juilland's frequency dictionaries, and cautions against an Anglo-centric account of the history.

Chapter 5: Corpus-Based Studies of Synchronic and Diachronic Variation

The chapter looks at variation across time and across contexts of use. For diachronic variation it considers language change from Old English to Modern English and in contemporary English, and for synchronic variation it contrasts the multi-dimensional approach associated with Douglas Biber, which looks at variation across genre and register using the text as unit, with variationist sociolinguistics, which looks at variation across social categories using the speaker as unit.

Chapter 6: Neo-Firthian Corpus Linguistics

The chapter examines the school of corpus linguists who follow J. R. Firth's ideas, led by John Sinclair at the University of Birmingham and including Michael Hoey, Susan Hunston, Michael Stubbs and others. Collocation and discourse are the two central ideas, and the chapter discusses how these terms are used.

Chapters 7–8: Corpus Methods and Functionalist Linguistics; The Convergence of Corpus Linguistics, Psycholinguistics and Functionalist Linguistics

Chapter 7 argues that corpus linguistics is converging with functionalist (cognitive or usage-based) linguistics: functionalists increasingly use corpus evidence, and some corpus linguists, notably Stefan Gries, use functional and cognitive theory to interpret results. Chapter 8 extends the argument to experimental psycholinguistics and discusses why neo-Firthians, who favour corpus-driven work without outside theory, relate to this convergence differently.

Chapter 9: Conclusion

The final chapter reflects on the place of corpus linguistics in linguistics. It describes two phases, an early period up to about the end of the 1980s in which the field established itself against Chomskyan scepticism, and a later one in which corpus methods became part of the toolbox of linguistics generally. It considers methodological pluralism and the potential for corpus methods in the humanities and social sciences.

Key Concepts

Chapters 1–3: Foundations and Practice

• Corpus: a set of machine-readable texts, usually of a size too large for manual analysis, chosen as a basis for particular research questions.
• Corpus as method versus corpus as theory: the distinction between "methodologist" corpus linguists and the neo-Firthian view.
• Introspection versus observed data: the shift in what counts as evidence in linguistics.
• Legal and ethical issues: the considerations in building corpora, especially from the web.

Chapters 4–6: Traditions and Variation

• Diachronic and synchronic variation: change over time and difference across contexts of use.
• Multi-dimensional analysis: Biber's approach to variation across registers.
• Collocation and discourse: the two ideas at the centre of neo-Firthian work.

Chapters 7–9: Convergence and Outlook

• Usage-based and functionalist linguistics: frameworks that treat language structure as shaped by use.
• Methodological pluralism: the authors' view of the future of the field.

Critical Analysis

Strengths

• A distinctive scope. The preface positions the book against introductory how-to textbooks: it deals with history, theory and the practical problems of research, so it fills a place those books do not.
• Both schools are presented. The book sets the corpus-as-method and neo-Firthian, corpus-as-theory positions side by side, and Chapters 7 and 8 trace how each relates to functionalist linguistics and psycholinguistics.
• Coverage of neglected issues. A whole chapter is given to legal and ethical questions, which the chapter itself calls less often discussed, including compiling corpora from the web.
• Reader support and reception. The publisher lists practical tasks, discussion questions at the end of each chapter and an extensive glossary. Stefan Gries (University of California, Santa Barbara) calls it a comprehensive and up-to-date overview covering theoretical debates as well as practical questions, and Karin Aijmer (University of Gothenburg) calls it a unique resource, in endorsements quoted on the publisher's page.

Limitations

• Not a procedural manual. By the authors' own statement it gives no step-by-step instructions and does not cover concordances, collocations and keywords in detail, so a student has to learn the tools elsewhere.
• English-centred history. Chapter 4 and much of the book's story follow English-language work, although the authors caution against an Anglo-centric account.
• Western standpoint on law and ethics. Chapter 3 says it is written from a Western standpoint and does not survey other jurisdictions.

Where It Fits in the Course

This is the course's method-and-theory text for corpus linguistics. Csomay and Crawford's Doing Corpus Linguistics is a step-by-step introduction with small tasks using public corpora, and is the natural first book for the hands-on side that McEnery and Hardie deliberately leave out. Read it first or in parallel if you have never searched a corpus.

For programming, Gries's Quantitative Corpus Linguistics with R teaches writing your own programs and statistics, while Bird, Klein and Loper's Natural Language Processing with Python and Jurafsky and Martin's Speech and Language Processing move into computational linguistics and language technology. Language and Computers, the open-access introduction, gives a gentler overview of language technology. McEnery and Hardie are the book to read for the questions behind these tools: what a corpus is, what counts as evidence and how the field fits into linguistics.

How to Use This Book

The book suits students who have had a first course in linguistics, upper-level undergraduates and graduate students designing a corpus project or writing about corpus methods. The preface presents it as different from introductory textbooks, so readers with no practical experience should pair it with a hands-on text.

A practical plan: read Chapters 1 and 2 first for the definition of the field and for how corpus data are accessed and analysed, then Chapter 3 before collecting any web data. Chapters 4 to 6 provide the history and the main traditions; read Chapter 6 on neo-Firthian work if your interest is lexis or discourse, and Chapter 5 if it is language variation or change. Chapters 7 and 8 matter most to students interested in theory or psycholinguistics, and Chapter 9 gives the authors' view of where the field is going. Use the practical tasks and discussion questions at the end of each chapter to test your understanding, and the glossary to check terms.