Katherine Bode has written a bold and illuminating book. A World of Fiction: Digital Collections and the Future of Literary History asks difficult questions and is groundbreaking in its interpretations. Treating Australian literature with tools and insights drawn from digital humanities, her work has important implications for literary study writ large.
Bode began with a simple inquiry: what fiction was published in nineteenth-century Australian newspapers? To better understand a past century she turns to a twenty-first century achievement, the Trove database created by the National Library of Australia. A world-leading resource, the Trove collection of digitised nineteenth- and early twentieth-century newspapers exceeds in both thoroughness and quality comparable efforts in Great Britain, Europe, and the United States. Having Trove enables Bode to explore newspaper fiction – most of it written from 1860–1900 – in a way never possible before. However, she does not merely accept Trove as given but instead actively improves upon it. In fact, she applies the word ‘edition’ to what she derives and develops: ‘for this book I did not simply theorize a scholarly edition; I built one. Its curated dataset is a subset of the 16,500 titles I discovered in the nineteenth-century Australian newspapers digitised by Trove’ (7). (Even after submitting her book manuscript, Bode has remained engaged in the search for fiction; her web-accessible dataset ‘To be continued . . .’: The Australian Newspaper Fiction Database has expanded to over 21,000 items at the time of this writing.) Bode breaks new ground in applying the principles of editing and textual scholarship to the analysis of a large scale literary system.
Initially, she regarded the compilation and refinement of her dataset to be precritical activity necessary to answer the synoptic questions of literary history she wished to address, but she quickly recognised the interpretive dimensions of corpus creation. Her study encompasses both canonical and noncanonical fiction and includes writing from everywhere around the globe, from Austria to Japan, from South Africa to Hungary. Not surprisingly, however, most of the writing comes from Britain, the United States and what is now Australia. Along the way, Bode has made numerous breakthroughs, including, for example, discovering previously undocumented fiction by Jessie Mabel Waterhouse and Catherine Martin, and she has located a ‘new’ Australian author, John Silvester Nottage who wrote several novels and numerous shorter works – thirty-two titles in all. Further, she has determined that ‘Captain Lacie’ and ‘Ivan Dexter’ were pseudonyms for James Joseph Wright who stands now as perhaps the most frequently published early Australian author.
As mentioned, she ‘draw[s] on the theoretical and practical foundations of textual scholarship’ to reconsider the level of rigour required if we are to make sound arguments about a literary system as a whole (4). Her efforts are designed to remedy how literary scholars use big data – or at least to provide a striking alternative to dominant practices. She typically manages to blend a prudent awareness of the limits of her own dataset with an ability nonetheless to make large and bracing claims. Occasionally her claims may strike some as excessive: in one paragraph she asserts that the ‘scholarly edition of a literary system can revolutionize knowledge of literary history’, and she describes ‘an entirely new organization for and structures within literary culture in the colonies’ (4, emphases added).
She contends rightly that a scholarly edition is an ‘argument – a historical and critical but also a technical one . . . It presents that argument through a curated text’ (6–7). She recognises that she is modelling literature based on an extensive but still partial representation of past production rather than on a complete record. Bode engages in ‘extensive historical bibliographical research and representation’ of the constituent works of a literary system and makes her own ‘rigorous engagement with a mass digitized collection’ available to all via the National Library of Australia (7–8). In making her data accessible, she follows what she persuasively argues should be core ‘methodological imperatives of quantitative literary scholarship: openness, testability and accountability’ (Bode, ‘Data’).
Newspapers were a hugely important venue for fiction. Bode analyses 313 newspapers and finds that 258 of them, or 82%, published extended fiction. Her findings overturn many assumptions. For example, she observes that Aboriginal characters have a much greater presence in early Australian writing than literary historians typically assert. She also demonstrates that Tillotson’s Fiction Bureau, a British syndication agency, was far less dominant than has been claimed. In fact, a local syndication service, Cameron, Laing, and Co., published the most Australian fiction in the nineteenth century. Bode also highlights the importance of provincial papers, demonstrating that they were more involved in reprinting and publishing fiction than metropolitan ones. She argues against the view that the relationship of Australian to British literature was one of subservience and imitation, finding ‘thematic tendencies in colonial fiction that are clearly different from those in American and British writing.’ American fiction was frequently labelled as British, pointing to the prestige value of Britain.
In the course of her work, Bode discovers a vastly enlarged record of Australian literary culture and its roles in the international circulation of fiction in the nineteenth-century. Bode practices ‘data-rich literary history’ with revisionary purposes in mind. Her work diverges sharply from the much-discussed methods pioneered by Franco Moretti and Matt Jockers. In a trenchant analysis, she finds their work to be inattentive to the historical complexity of textual conditions. For example, they rarely if ever acknowledge that the status and reliability of any particular version of a work is contested. She likens their practice to that of New Criticism: in a detailed and nuanced argument, she highlights their lack of attention to historical context, to textual multiplicity and fluidity, and to the history of publication and reception.
A refreshing aspect of this book is Bode’s engagement with ‘pressing questions about the nature and implications of literary history conducted with mass digitized collections and the literary data derived from them’ (3). In a nutshell, her goal, informed by rich data, is to develop a noncanonical and transnational history of the literary publishing and reading cultures of nineteenth-century Australia. Working with Trove, however, complicated matters, opening up new and vital questions about literary history based on mass-digitised collections. She became increasingly aware of the gaps in Trove, even though this newspaper digitisation project is one of the most capacious in the world. Her commendable caution sets her apart from critics who sometimes speak as if their data set captures the totality of past writing. She is well aware that the Optical Character Regognition (OCR) underlying Trove is riddled with errors, and, despite extensive crowd sourcing efforts to the clean the data, it remains problematic. Yet Trove has a crucial advantage over other comparable digitisation projects in that it is organised at the article level (rather than at the page level) and the title and first words have been keyed and checked resulting in a high degree of accuracy. Relying heavily on titles, paratextual clues, and the opening words of individual items, Bode is able to identify a vast amount of fiction.
A key argument of this book is that those involved in data mining or text mining need to reconsider their methods. Bode calls for an approach that accommodates anonymous and pseudonymous works. We need a much better understanding of ‘unknown’ or ‘anon’ given that more than a third of the fiction in this period was published without authorial attribution. As she notes, if we lessen our emphasis on authorship we gain the benefit of intensified focus on reception and can offset the ‘disproportionate power of contemporary reputation created by literary canons’ (93). We can also avoid the distortion created by the enduring power of the nation to organise and constrict our view of literary publishing and reading culture.
Acutely aware that corpus creation is its own scholarly activity, Katherine Bode provides a combination of attentiveness to detail and a capacity for large arguments. The results are powerful: her scholarship is as expansive in scope as it is scrupulous in method. A World of Fiction makes a landmark contribution to literary studies and digital humanities.