In research on academic discourse, there is an extensive literature devoted to the analysis of research articles (RAs), much of which focuses upon determining the characteristic features of disciplinary discourse. However, as suggested by its subtitle, Bethany Gray's Linguistic variation in research articles: When discipline tells only part of the story offers a new and more fine-grained analysis, which reveals discipline to be only one of the important factors accounting for the variation found in RAs. Thus Gray's aim is not just to determine the linguistic features that characterise RAs across different disciplines, but also to investigate and describe how RAs vary within disciplines.
This book consists of eight chapters, along with six appendices, containing further statistical data. After the introduction in chapter 1, Gray devotes the next two chapters to setting out the methodology of the study, with chapter 2 presenting the procedures used for establishing a taxonomy of RAs, while chapter 3 describes the construction of the Academic Journal Register Corpus used in the analysis. Chapters 4–7 are the core of the book, offering three different types of analysis: first comes a situational approach in chapter 4, followed by two chapters which deal with lexicogrammar. Chapter 5 presents a general lexical and grammatical analysis, while chapter 6 focuses specifically on the issue of structural complexity in RAs. A multidimensional analysis is provided in chapter 7, before the outcomes and implications of the research are summed up in chapter 8.
In chapter 1, ‘Introduction’, Gray surveys the literature, focusing first on the linguistic features of academic writing in general and then more specifically on the RA itself. She notes that the RA is probably the most researched academic register and that analyses primarily concern aspects of lexis, grammatical structure and generic organisation. In summarising relevant studies on RAs, she presents a useful table, which gives information on the disciplines, linguistic features and comparisons with other registers covered by each study. She points out, however, that despite this considerable body of work, little account has been taken of variation within the RA register and there is a lack of systematic and large-scale comparisons of core linguistic features across disciplines. Her book, then, aims to remedy these shortcomings and the chapter concludes by setting out the goals of the study and explaining the overall approach.
Chapter 2 seeks to determine the different types of texts that are prevalent in the domain of academic journal writing. In order to do so, texts from a range of journals in eleven disciplines are examined, which leads Gray to establish three overarching article types: empirical, theoretical and evaluative. Each of these is further classified into three or four subtypes; for example, empirical articles can be qualitative, quantitative experimental, quantitative observational or mixed methods. For each subtype Gray gives definitions based on both content and textual/genre features, thus setting up a useful taxonomy of article types, which should enable analysts to classify the texts in academic journals more rigorously than before. The chapter goes on to examine the extent to which each subtype of article is prevalent in the disciplines analysed. This not only reveals wide variation within the disciplines, but also shows that no discipline makes use of all article subtypes. Gray uses this analysis as the basis for selecting the disciplines and subtypes to be included in the Academic Journal Register Corpus, which is described in the following chapter.
The corpus consists of articles in three journal registers (theoretical, quantitative and qualitative) and six disciplines: philosophy, history, political science, applied linguistics, biology and physics. Where possible, disciplines that make use of two registers are chosen to enable comparisons within disciplines to be made. Thus applied linguistics and political science are represented by both qualitative and quantitative articles, while the physics component contains both theoretical and quantitative studies. Chapter 3 describes how the corpus was constructed, drawing not only on the results of the journal survey and taxonomy, but also on interviews with specialist informants to inform the choice of articles for each discipline and register. The corpus consists of 270 RAs, 30 from each discipline/register category, taken from 56 journals and amounting to about 2 million words and is automatically tagged for part of speech.
In chapter 4, Gray establishes a framework which she uses to analyse the situational variation present in the RAs in the corpus. The framework consists of eight categories of characteristics: participants, textual layout and organisation, setting, subject/topic, purpose, nature of data or evidence, methodology and explicitness of research design. Each of these is explained and discussed in individual sections, before Gray goes on to characterise and exemplify each discipline/register category according to its use of the situational features in the framework. She points out several interesting tendencies revealed by this analysis and discusses possible reasons for them. For example, the number of authors is closely related to discipline and register, with qualitative history and theoretical philosophy papers tending to have single authors, while quantitative biology and physics papers are usually multiple-authored. She also notes that papers in the hard sciences do not tend to state their research questions explicitly, in contrast to those in the social sciences. The wealth of data in this chapter is summarised in the form of a table which allows the reader to compare disciplines and registers with ease.
Chapter 5 presents a lexical and grammatical survey of the disciplines and registers, focusing on core grammatical features and including individual sections on nouns, pronouns, verbs and verb phrases. Nouns and verbs are analysed into semantic sets and verbs are characterised by tense, aspect and voice. Tables show the distribution of individual features across disciplines and registers, thus revealing both similarities and differences in their use. Certain findings are as expected, for example, passive voice occurs more frequently in the natural sciences than in the other disciplines. However, Gray's data also show that there is variation within disciplines according to register. For example, the occurrence of passive voice in quantitative applied linguistics resembles that in other quantitative disciplines and is considerably higher than in qualitative applied linguistics.
In chapter 6, Gray turns her attention to the issue of structural complexity in the corpus. She builds upon earlier work which characterises academic writing as structurally compressed, with a high use of noun phrases, in contrast to conversation, which tends towards structural elaboration (Biber & Gray Reference Biber and Gray2010). In the present volume, Gray examines features of clausal elaboration as well as phrasal compression and shows that their use varies across disciplines and registers, with humanities disciplines having a generally higher use of clausal elaboration, while the natural sciences tend to privilege features of phrasal compression. However, in all disciplines there is a striking difference between the prevalence of compression and the much lower incidence of elaboration.
In the last of these individual investigations, in chapter 7, Gray provides a multidimensional analysis of the corpus, an approach in which clusters of co-occurring features are determined statistically, allowing the analyst to situate and group texts along several dimensions. Four dimensions are distinguished as important in characterising the RAs in this corpus: (1) academic involvement and elaboration vs information density; (2) contextualized narration vs procedural discourse; (3) human vs non-human focus; (4) ‘academese’. Gray notes that although these texts all belong to the same broad register of written academic discourse, the first dimension (involvement/elaboration vs information density) nonetheless distinguishes between them. Thus theoretical philosophy shows more features of involvement/elaboration than all the other disciplines, while quantitative natural sciences tend to use more features of information density than quantitative social sciences. She argues that this dimension identifies the features of grammatical complexity that characterise the register of professional academic writing.
Chapter 8 draws the outcomes of the analyses together and links the linguistic findings to the situational analysis offered in chapter 4. These linkages are succinctly presented in table 8.1, which shows how the prevalent linguistic features of each discipline can be related to characteristic features of its situational context. Gray makes the further point that variation in the occurrence of these features occurs on a cline from soft to hard disciplines. In terms of variation within disciplines, Gray finds, for example, that dimension 2 of the multidimensional analysis distinguishes between qualitative and quantitative articles, with qualitative research tending towards contextualized narration, while quantitative studies make more use of features of procedural discourse. However, Gray stresses the multilayered nature of academic writing and sees the variation in these articles as the product of a complex interplay between register and discipline. The book closes by making recommendations for future corpus design and offering suggestions for further investigations in this area; in particular, Gray advocates the use of complementary approaches to the analysis of variation in academic discourse.
In evaluating this book, I would first like to stress that it makes a very important contribution to the study of academic discourse and should be required reading for all who are concerned with researching and/or teaching the RA. Nonetheless, it is also necessary to draw attention to certain areas of weakness. The first of these concerns the selection of disciplines to analyse. While I acknowledge the careful way in which the disciplines have been chosen for the purposes of this study, it is still disappointing that there is no detailed analysis of an applied natural science (e.g. pharmacy) and that one of the social sciences analysed is applied linguistics. Much analysis of this latter discipline is already available and it would have been good to see another less studied discipline (e.g. social work) represented.
The second area to mention is the coverage of the literature. Clearly there is always a time lag between research and publication so that it is not possible for a literature review to be absolutely up to date. However, this does not account for the fact that some important studies are missing from the discussion. I will cite here just two that particularly caught my attention. Thus the early work of Becher (e.g. Becher Reference Becher1981) is referenced, but not the later volume co-written with Trowler (Becher & Trowler Reference Becher and Trowler2001), which gives a much more detailed analysis of disciplinary cultures. Similarly, although Swales’ (Reference Swales1990) work on RAs is referred to, there is no mention of his (Reference Swales2004) volume on Research genres, which updates the earlier analysis. However, this book should not be judged on the coverage of its literature review and while one might have liked a rather more comprehensive treatment, the survey is adequate to the task in hand, which is to situate the author's present work in its research context.
Finally, it is a pity that the volume is marred by some careless editing. This consists not just in simple proofreading mistakes, which occur sporadically throughout the text, but also in some problems with the content. For example, in table 7.1 on page 140, item 54, ‘that-clause controlled by factive (certainty) adjective’, is exemplified by ‘conclude’ and ‘proves’, which are verbs not adjectives. A further example occurs on page 108, where discussion of the use of ‘we’ is followed by examples 5.83–5.85, which contain only the pronoun ‘I’.
However, these are relatively minor quibbles in comparison with the strengths of this book, which lie in its careful and painstaking analyses, supported by clear and convincing evidence of the subtle variation in RAs. Gray is to be applauded for her groundbreaking work and the volume is certain to become an established reference for all who are interested and involved in the analysis of written academic discourse. I myself will certainly be referring to it repeatedly in my teaching and research.
