Reading Numbers by Numbers: Digital Studies and the Victorian Serial Novel

Reading Numbers by Numbers: Digital Studies and the Victorian Serial Novel Susan David Bernstein (bio) and Catherine Derose (bio) Introduction With the plethora of formats for reading that digital technology has unleashed in recent years, from iPads and Kindles to blogs and Twitter, both reading and writing have undergone a sea change. In the middle of the nineteenth century, there was likewise a major transformation in the tempo of reading and writing as literally hundreds of weekly and monthly periodicals were launched. For most Victorian novelists, including Charles Dickens, Elizabeth Gaskell, Anthony Trollope, and George Eliot, serial instalments shaped both how they wrote fiction and how their novels were first encountered by readers. Typically, we think of the Victorian novel as a tome requiring huge blocks of reading time. But these novels were disseminated in segments of a few chapters, something like twenty or thirty pages of text, usually, but not only, in weekly or monthly instalments. Mark Turner has offered several temporalities of reading and writing generated by Victorian periodicals, from dailiness and weekliness to monthliness, quarterliness, and annualness (189). Even more intriguing for Turner is the disruption of time, the pause in the duration of story enforced by the very nature of serial publication: “In the breaks in the narratives of periodicals and in the lapses in time—over a day, over a week, over a month—is where meaning resides” (194). Although Turner addresses the publication of periodical magazines here and refers to the interactions of readers who wait for the next issue, we reorient this idea of the meaningful pause between parts to the serial novel. If not the volume or the chapter, what does the serial number, that unit of text through which Dickens developed his bulky multiplot novels, signify? What observations can we cull from a critical analysis of these parts in relation to the greater whole of a novel written and read over nineteen months, the usual schedule of Dickens’s part-number issue of novels, including Martin Chuzzlewit, David Copperfield, Bleak House, Little Dorrit, and Our Mutual Friend? Put differently, what happens when we think about novels as assembled pieces of published chunks attuned to the rhythms of seriality? The objective of this study is to detect, describe, and reread the signal of seriality in novels by Dickens and Eliot using a computer-assisted text analysis program called DocuScope, among other statistical apparatuses. With the assistance of digital tools, we [End Page 43] identify and analyze previously underexamined rhetorical shifts in serial novels, such as how setting and character distinctions separate Dickens’s weekly and monthly instalments, how temporal certainty is reflected in Eliot’s serialized and non-serialized novels, and the degree of references to the materiality of language that distinguishes Dickens’s and Eliot’s serials from the latter’s non-serialized novels. A juxtaposition of computer- and human-generated analyses demonstrates the potential of algorithmic criticism to reveal new areas of investigation. Docuscope Overview DocuScope was first developed at Carnegie Mellon University in the late 1990s by David Kaufer and Suguru Ishizaki. The program aims to help readers see representational patterns in texts; it was originally designed to provide fast statistical comparisons of student papers for composition/rhetoric classes. As text-tagging and visualization software, DocuScope behaves like a “smart dictionary” (Hope and Witmore 365), sorting electronically formatted texts for predetermined linguistic items (words and strings of words) that have been threaded to specific rhetorical categories (such as FirstPerson, Immediacy, and ReportingProcess). DocuScope contains three sifting categories that fold one into the other like nesting eggs: seventeen clusters, fifty-one dimensions, and 101 language-action types (lats). For example, the cluster FirstPersonInterior houses the dimensions FirstPerson and FirstPersonInterior, with the former dimension containing the lat FirstPerson and the latter one including Self-Disclosure, Self-Reluctance, and Autobiography. In our study, we focused on the 101 lats, which provide a finer-grained analysis than the two umbrella categories. The identified word strings and the categories to which they are assigned are the result of human input; Kaufer handcrafted DocuScope’s default dictionary using the Oxford English Dictionary as a template (Hope and Witmore 365). He then cross-referenced DocuScope’s...

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.