Abstract
The burgeoning community of academic research has seen an exponential growth of studies exploring the ‘two pillars of language research’ (Sinclair, 2004: 11): corpus linguistics and discourse analysis. Taking a unique view on what has been missing, the volume under review provides a timely and invaluable critical reflection on overlooked or under-researched areas and offers useful approaches to identify blind spots and avoid pitfalls in corpus-based discourse studies. As such, this volume contributes to enhancing the completeness of corpus-based discourse analysis and signifies ‘an important milestone in the development of corpus linguistics methods’ (Baker, Chapter 13, p. 291).
In the introductory chapter, Marchi and Taylor set the scene by distinguishing between qualitative and quantitative analyses and addressing the partiality in corpus-based discourse studies. The subsequent 12 main chapters revolve around the following three themes: (1) checking neglected aspects, under-researched topics or text types; (2) identifying undetected or under-analyzed aspects using triangulation and (3) re-examining some corpus methodologies in discourse analysis.
Part A accommodates overlooked aspects in the research agenda of corpus-based discourse analysis, in terms of similarity (Chapter 2), absence (Chapter 3) and overlooked text types (Chapter 4). With a purpose to check the overlooked areas, Taylor in Chapter 2 introduces some useful methods for similarity analysis and showcases a similarity analysis of expressions collocated with refugees in political and media discourse corpora. This perspective allows researchers to attain a ‘360 degree view’ (p. 35) of the discourse features under scrutiny, thus providing a complementary focus for variation analysis. Duguid and Partington (Chapter 3) report three cases of identifying salient absence in diachronic corpora, which underscores the importance of choosing or designing suitable comparison corpora and using appropriate comparison methods. In Chapter 4, Lischinsky investigates discursive construction of the gendered body in erotic fiction, which provides evidence for the complex relationship between language use and its social discourse.
Part B concentrates on identifying blind spots in corpus-based discourse analysis by using triangulation. In Chapter 5, Caple investigates multimodal discourse features in Instagram posts on the Australian federal election. Her analysis demonstrates the value of introducing nonverbal elements in analyzing meaning construction in media discourse. From a vantage point of triangulating multiple data, Jaworska and Kinloch (Chapter 6) cross-check postnatal depression across four datasets. The focus on the shared keywords in the four datasets reveals some commonality of the collocational patterns of depression. However, a minor reservation is that this focus may minimize the chance to obtain a comprehensive picture regarding distinctive features that do not appear throughout all the four datasets. Taking an interdisciplinary perspective, Ancarno (Chapter 7) details a study involving the collaboration between corpus linguistics and anthropology in analyzing 3-grams and their concordance lines.
Switching to methodological choices, Part C re-examines some fundamental corpus approaches in corpus-based research design. In Chapter 8, Egbert and Schnur provide an in-depth discussion of the roles of texts in corpus-based discourse analysis and showcase the roles via a case study of keyword analysis. Their study emphasizes the importance of using texts as a sampling unit and observing unit in corpus-based study. Following an extensive exploration of time division approaches, Marchi (Chapter 9) advocates adopting flexible, data-driven and multipurpose-based time division approaches in data segmentation to meet particular research purposes.
In Chapter 10, Anthony showcases some useful visualization methods for presenting quantitative corpus data. His critical evaluation of strengths and weaknesses provides a useful methodological reference for researchers seeking to visualize their discourse data. Chapter 11 contains a detailed account of keyness analysis and the metrics used to identify key items. Here, Gabrielatos also showcases how to use principled techniques to carry out keyness analysis in examining the differences and similarities in two UK election manifesto corpora. In Chapter 12, Brezina proposes a three-step analytical procedure and outlines several statistical options for cross-corpora comparison. His critical analysis helps to enhance researchers’ statistical literacy and enables them to make informed choices regarding statistical techniques in corpus-based discourse. In the concluding chapter, Baker recaps some key points in this volume and guides us through an in-depth rumination on partiality and reflectivity in corpus-based analysis. Most notably, he draws on Burr’s (1995) action research and proposes to use discourse findings to make an impact on social reality and ‘inspire some sort of social change’ (p. 285).
Overall, this volume brings to light many overlooked or under-researched areas in corpus-based studies and showcases some useful corpus tools for discourse analysis. It reminds us to take a balanced perspective and keep open to potential overlooked areas in discourse analysis. Both experienced and novice researchers involved in corpus-based discourse analysis may find this book a useful resource to raise their awareness of possible pitfalls, enhance their methodological choices and sharpen their discourse analytical tools.
Nevertheless, there is still room for minor improvement. The quantitative analyses in this volume (as well as in most corpus-based discourse studies) have predominantly focused on the overall frequency of discourse features. However, another important aspect, concerning the textual colligation, namely at what textual position a discourse feature often appears, remains under-researched. As Dong and Buckingham (2018: 432) have found, textual positions represent ‘the timing with which linguistic units are deployed to achieve certain communicative purposes’; thus, future studies may consider incorporating this aspect to gain new insights into discourse features as well as to supplement the research arena in discourse analysis.