Abstract
Mahlberg’s book is, as the title suggests, a stylistic analysis of Dickens’s published works. The two stated aims of her book are ‘to outline a corpus stylistic foundation that enables the use of innovative descriptive tools for the analysis of literary texts’ (p. 3) and then to apply those tools to Dickens’s works and reveal the ‘textual building blocks of fictional worlds’ (p. 3).
Mahlberg argues that a corpus stylistic analytic approach is a needed addition to both the existing body of analytic and interpretive work about Charles Dickens’s writings, and also to literary analysis more generally. She acknowledges that a large volume of literary analysis of Dickens’s work already exists. However, through the power of computer analysis, researchers are able to quickly and efficiently detect linguistic features within one text from the body of an author’s work, and from works of a certain time period. Researchers can then consider the function of those features in the creation of the fictional worlds of one author and, moreover, how typical or atypical they were for the time in which a text was published (p. 2).
The book comprises an introduction, nine chapters and nine appendices. The first four chapters explain the details of Mahlberg’s methodology and include an outline of the relationship between corpus linguistics and literary stylistics. Chapters 5–8 provide the results of her analyses and interpretation of the data. The final chapter concludes her analysis and outlines several directions for future research. The appendices contain large data sets from which she draws in Chapters 5–8.
Mahlberg begins by explaining how corpus analysis can contribute to literary researchers’ and theorists’ knowledge of linguistic patterns and literary interpretation. Interestingly, she acknowledges the important role of actual readers, and states that she intends to ‘explore the relationship between linguistic units and the contributions they might make to the effects that texts have on readers’ (p. 6, my italics; see also pp. 32–39). Moreover she frequently refers to ‘the meanings of items in texts’ (p. 17, my italics). Disappointingly for this reviewer, Mahlberg’s stated interest in readers and meanings did not result in her creating data about how actual readers made actual meanings with, or sense of, the linguistic patterns she extracted from Dickens’s works.
In Chapters 2, 3 and 4, Mahlberg explains that ‘clusters’ of words are ‘textual patterns that are relevant to the creation of characters in readers’ minds’ (p. 26). Examples of such clusters are: ‘dear Copperfield said Mr Micawber’, ‘upon my word and honour’ and ‘not a bit of it’. Mahlberg describes and explains in detail the criteria she set for deciding on the number of words in a ‘bundle’ or ‘cluster’ (Chapter 3), and then (Chapter 4) how she categorized the clusters into groups (e.g. Body Language) which form ‘the basis for a description of areas of meanings and functions’ (p. 41).
Mahlberg details her findings in relation to the clusters of Characters’ Speech (Chapter 5), Body Language (Chapter 6) and Labels (Chapter 8). In Chapter 7 she deviates from cluster analysis by using concordancing data to examine a cluster she labelled ‘As If’. In each chapter Mahlberg explains what kinds of linguistic aspects each cluster has (defining features) and some existing theories about those features (e.g. in Chapter 5 she discusses aspects of pragmatics and discourse analysis), and then analyses and discusses those features in the context of the texts. She argues that with all the clusters, a fruitful analytical approach is to place them on a continuum or ‘cline’ of ‘contextualising and highlighting functions’. At the ‘contextualising’ end the function is to see the ‘information’ in the cluster as subtle and best interpreted as part of the larger textual picture. At the ‘highlighting’ end, the cluster stands out because it is in some way striking for a reader (pp. 82–83).
Although Mahlberg has written clearly, explained well and been forthright in her wish not to over-interpret the data, this reviewer couldn’t help but feel that her constant references to ‘readers’ called for data from actual readers. Mahlberg does make a reference to this kind of data in her concluding Chapter 9, suggesting that a psycholinguistic investigation might fruitfully investigate to what extent readers are ‘actively aware of the repetitions of clusters and how noticeable really is the difference between highlighting and contextualising clusters’ (p. 177).
Actual readers might also have provided another view on the data which Mahlberg seems to have overlooked in her reliance on computer analysis. For example, the identification of Body Part clusters relied upon body part nouns being identified in the corpora (e.g. ‘his hands in his pockets’). However, references to body parts may also be made through anaphoric pronominal reference. For example, it is possible that the hands were, in adjacent sentences, referred to again using a pronoun. These additional instances of the cluster were not identified, thus reducing the numerical weight of the body part. Although researchers could simply extend the computer program to take account of that apparent oversight, it is not possible for a computer to read like a human. To paraphrase Louise Rosenblatt (1994: 10), readers don’t finish reading the first page until they have finished the last page. This affects Mahlberg’s cline of ‘contextualising and highlighting functions’ in that where an item might be on the cline depends on when the reader is making sense of the item. And in recursive, iterative reading the reader is potentially frequently making multiple senses of an item – it might have multiple positions on the cline.
Notwithstanding those comments, Mahlberg’s book will be of interest to postgraduate students in corpus linguistics and literary studies. It is clearly and engagingly written and has insights to offer those who want to use computers to analyse texts.