A title in the Cambridge Elements (pragmatics) series, Corpus Pragmatics is one of the most recent contributions to this well-established field that has its own dedicated journal, and a number of handbooks, textbooks and edited volumes, but remains both vibrant and rapidly evolving. Corpus pragmatics is an approach to the study of pragmatics that focuses on the use of corpora and associated analytical software for a quantitative and qualitative interpretation of language in use. Corpus pragmatics as a distinct approach first came to the fore in Romero-Trillo’s (Reference Romero-Trillo2008) edited volume, which successfully brought together previous research that had not explicitly used the term while paving the way forward for the field. Landert, Dayter, Messerli and Locher’s volume distinguishes itself from previous work by exclusively focusing on a function-to-form approach. Traditionally, corpus pragmatics has often prioritised form over function and the authors seek to address this. Usefully, each chapter contains an illustrative case study demonstrating real-world applications of this methodological focus.
Chapter 1 (‘Introduction’, pp. 1–9) outlines the challenges faced by researchers using corpora to study pragmatic phenomena. These include lack of access to any meaningful (extended) context, the privileging of quantitative over qualitative results and a focus on form over function. The chapter highlights one of the primary concerns of the volume – to what extent the identification of pragmatic functions in future studies will be possible without recourse to either pragmatic form or time-consuming manual annotation or interpretation. The chapter ends on a more positive note for those interested in corpus pragmatics, as the authors emphasise the opportunities and benefits offered by a corpus approach.
Chapter 2 (‘Corpora and their characteristics: Challenges and opportunities for pragmatic research’, pp. 9–20) continues on the positive closing note of chapter 1. It showcases a wide variety of different types of corpora, many of which are freely available to the researcher. Nine different types of corpora are outlined and a description provided for each. The advantages and limitations of using a particular corpus type are discussed and illustrated using previous research. This previous research highlights the breadth of corpus pragmatics, which demonstrates that it is ‘far from a homogenous field of research’ (p. 20). The authors stress that every corpus type presents its own set of challenges and opportunities. The analyses of different corpus types require an array of different skills – for example, annotating a small corpus is a very different task from working with large datasets unsuited to such annotation.
The use of a self-compiled corpus for pragmatic research is explored in detail in the volume. Chapter 3 (‘Corpus-driven approaches with self-compiled corpora’, pp. 20–7) begins with the distinction between corpus-based and corpus-driven approaches. A corpus-driven approach is one that is relatively free of preconceived notions about language as the corpus itself creates the research hypothesis. The authors suggest that corpus-driven pragmatic research requires relatively large corpora and sophisticated automatic annotation to avoid researcher bias. A number of benefits and challenges of this approach to pragmatic research are discussed, and these can be usefully extended to other research contexts. The chapter points towards a number of issues that should be addressed, including corpus design, ethics and copyright, and technical challenges, through a case study centred around a corpus-driven study of the language of persuasion.
Chapter 4 (‘Corpus-assisted approaches with self-compiled corpora’, pp. 27–39) discusses perhaps the key issue in a function-to-form approach: how the gap between the automated analysis of large-scale corpora and the close reading of a corpus by the researcher can be bridged. Bridging the gap is important as this integration of automation and close reading facilitates an empirical, unbiased, quantifiable and reproducible corpus process that eliminates the common criticism of ‘cherry picking’. The corpus-assisted approach to corpus pragmatics is illustrated by an in-depth look at the Language and Health Online project. Although the analytic process demonstrated is time-consuming, and only 10 per cent of the overall corpus is qualitatively analysed, the case study does show how the combination of qualitative and quantitative analysis allows for replicability and systematicity.
In a general sense, corpus linguistics has long been concerned with the challenges presented by working with pre-existing corpora and chapter 5 (‘Compatibility and comparability: Combining existing corpora’, pp. 39–47) explores this hoary old chestnut. The authors are keen to stress that there are very solid reasons for the use of pre-existing corpora in corpus pragmatic research, but caution that issues of compatibility and comparability should always be considered. They recommend a familiarisation process which involves using corpora locally and generating results on one chosen concordancer for comparative purposes. Frequency normalisation is also crucial and an excellent case is made for this to be done in the traditional way, without recourse to figures from concordancers, given the variation across different corpus interfaces even when performing a simple task such as identifying the number of words in a corpus. When these issues are considered, results produced by different corpora can be interpreted in a pragmatically meaningful way.
Chapter 6 (‘Scalability: Meaningful pragmatics with large data’, pp. 48–56) argues that one of the major challenges faced by the researcher is how to deal with the issue of time when confronted with large corpora. One of the ways that this is traditionally dealt with is to use a sample of the entire corpus, using random or stratified sampling. In both instances, the sampling unit and sample size should be considered for optimal efficiency. Two case studies, the speech act of self-praise by politicians and epistemic stance in Early Modern English, are used to demonstrate how to improve the scalability of qualitative analysis with large corpora.
Chapter 7 (‘Multimodality: Integrating non-verbal information’, pp. 56–63) is a timely chapter given the growing concerns of how to integrate different types of non-verbal information into a corpus, such as images, hand gestures or emojis. The chapter sets out the challenges for multimodal corpus pragmatic research. Primary amongst these is how to accurately represent acts of communication as they occur in the real world in a way that facilitates pragmatic study. The extension of this, then, is the process of encoding the information so that it is corpus readable. This is likely to involve the creation and development of new corpora, methods and software that can systematically process this information. The chapter also outlines the multimodal corpora that are freely available with the caveat that these are limited given the degree of expertise involved in building a corpus of this type.
Finally, chapter 8 (‘Open issues and outlook’, pp. 63–7) highlights what the authors believe needs to be addressed in the future to progress a corpus pragmatic research agenda, especially in relation to a function-to-form approach. The authors recommend the need for more open-access corpora, especially those specifically designed for pragmatic research. These corpora need to be designed and built with the provision of adequate contextual information in mind so that meaningful pragmatic interpretation can be performed. Corpus tools themselves also need to be reconsidered in light of a function-to-form approach. The development of corpus tools that facilitate large-scale qualitative analysis in pragmatically meaningful ways would be a boon.
This Cambridge Elements series is committed to showcasing dynamic, high-quality, original and concise work in pragmatics. This volume certainly covers these bases and is a highly informative read, despite the space restriction of the series in general. The field of corpus pragmatics is so dynamic that the volume, although published in 2023, predates advances in pragmatic annotation that may be possible using AI (see, for example, Yu et al. Reference Yu, Li, Su and Fuoli2024). These AI tools could, in time, potentially alleviate the arduous process of the pragmatic annotation of corpora raised in this volume. The Elements series also stresses accessibility and the volume is suitable for a postgraduate and expert audience, though because the volume presupposes knowledge of a wide array of associated computer software, it is perhaps a little too dense for undergraduates. Nonetheless, this is an excellent contribution to the field of corpus pragmatics and has already become a very useful resource for this researcher.
