Back to Book Reviews

Back to Cercles


Corpus Stylistics

Theory and Practice


Dan McIntyre and Brian Walker


Edinburgh: University Press, 2019

Paperback. xviii+356 p. ISBN 978-1474413213. £24.99


Reviewed by Laure Gardelle

Université Grenoble Alpes




This volume is intended as a practical guide to corpus stylistics – how to build and annotate a corpus, how to use existing corpora (such as the BNC and its BYU interface) and major software packages (such as Antconc or Wmatrix). It provides theoretical and methodological discussions, as well as many case studies, step-by-step guidelines and screenshots. Corpus stylistics is a recent research area (the collocation probably first appeared in print as recently as 2004), and not all stylisticians are convinced of the assets of a corpus approach. The aim is to show that new insights can be gained; in particular, intuitive claims about style can be tested. The authors also seek to show that corpus stylistics is a ‘distinctive analytical practice’ from corpus linguistics [21], and that the two can mutually benefit each other.

Because corpus stylistics is recent, existing definitions differ, ranging from a subfield to a methodology. To the authors, stylistics, ‘the linguistic study of style in language’, seeks to identify the linguistic properties of a given text-type (literary language most of the time, but other types as well); and corpus stylistics is ‘the application of theories, models and frameworks from stylistics in corpus analysis’ [15-16].

The book is divided into ten chapters. The first five focus on theoretical and methodological issues; chapters 6 to 9 then provide practical examples, including uses for L2 classroom activities; chapter 10 then rounds up the arguments, making a case for corpus stylistics as a useful development. Here is an overview of the topics broached in the volume, based on what the present reviewer sees as the most salient contributions.

A first technique for corpus analysis in stylistics is corpus-informed stylistics, that is, the use of linguistic corpora to establish a norm, from which the interpretation of a text may be carried out (chapter 2). One example is the study of collocational behaviour, which is responsible for expected schemas of meaning. It is illustrated by the study of with intent in the phrase ‘strolling with intent’ in a novel by John Le Carré. A search for ‘with intent’ in the BNC reveals that most uses are associated with law breaking; this high-frequency collocate, it is argued, influences the way the reader interprets strolling in the Le Carré quote. The present reviewer would suggest that this does not inform on style so much as on semantics and connotations for language users – the kind of information lexicographers derive from corpus analysis. Still, such research is useful to refine interpretation in context. Another instance of corpus use is semantic prosody: knowing about the collocates of a word may point to unprototypical combinations (such as utterly combined with a positive adjective), to provide evidence of irony or other stylistic effects. The authors lay out such important distinctions as ‘frequency’ vs. ‘collocational strength’, or ‘semantic prosody’ vs. ‘semantic preference’. They conclude convincingly that although corpus stylistics cannot bring all the answers, it does provide ‘a means of seeing patterns in data that would not be observable through purely qualitative analysis’ [63].

Corpus-based stylistics may also rely on specifically annotated corpora (chapter 3). The authors provide advice on how to design a balanced corpus and get a truly random sample using Excel; they also consider ethical issues and copyright. The discussion is very well-informed. Then hands-on advice is given to produce machine-readable text, split or merge data files, annotate the corpus automatically or manually. Again, the help boxes and case studies are very helpful.

Chapter 4 considers research questions and reliability of research findings, with the key notion of falsifiability. The chapter provides advice based on how to make generalisations about the style of an author, or characterise the style of a given text (keyness studies). As noted by the authors, a single text is not properly a ‘corpus’, but corpus-inspired methods give useful insights. The chapter also considers corpus-based and corpus-driven analyses, the role of intuition and hypotheses, and the issue of objectivity. Chapter 5 addresses the notion of statistical significance of raw figures, and proposes an overview of a number of core statistical tests. All the concepts, such as degree of freedom, are defined in accessible terms, which helps the reader make sense of what the calculations mean. The assets and limitations of the various tests are very well spelt out, and illustrated by a number of case studies.

Chapters 6 to 9 are devoted to full case studies, detailing all the steps from the aim of the study to research questions and how corpus analysis can help. The studies include real-world projects, from language consultancy for political organisations to assessment of the impact of intralingual subtitling (for deaf and hard-of-hearing viewers) on characterisation – in order to improve users’ experience of TV drama.

This volume is a must-read for beginners in corpus stylistics. It combines theoretical and methodological discussion with very practical advice. The step-by-step instructions, with screenshots, are very helpful, and the many case studies do prove that corpus stylistics is both possible and useful. Just as in linguistics, a corpus-based approach does not mean that qualitative analyses are rejected; rather, as the authors are committed to showing, corpora provide tools for new insights.



Cercles © 2020

All rights are reserved and no reproduction from this site for whatever purpose is permitted without the permission of the copyright owner.

Please contact us before using any material on this website.