Chapters and Digital Components

The Viral Texts project has been from its inception a fully interdisciplinary project: a collaboration among literary scholars, historians, computer scientists, and other experts from a range of fields, as well as between scholars at different ranks, including graduate and undergraduate student researchers, librarians, and disciplinary faculty. We are dedicated to reflecting the collaborative and interdisciplinary core of the project in Going the Rounds, and publish it as a multigraph, with Ryan Cordell as the primary author, co-writing each chapter or digital section with other members of the Viral Texts team. As a multigraph, the book develops a consistent set of arguments, like a monograph, but includes a wider range of writers’ voices that reflect the interdisciplinary work that underlies its arguments. We offer Going the Rounds as one model for academic authorship that rightly reflects the collaborative work increasingly common in humanities research projects.

As a result of the multigraph approach, you will find while reading this book that its style and focus varies somewhat from chapter to chapter, reflecting the disciplinary backgrounds and aims of each chapter’s authors. Some chapters foreground the literary and historical insights we glean from our data on nineteenth-century reprinting, for example, and are structured primarily toward humanistic argumentation, while other chapters focus on the computational methods we have developed and seek to intervene in critical debates in computer science, computational linguistics, and related fields. One of the most rewarding aspects of the Viral Texts project has been that our collaborations have proven generative across our fields of expertise. In writing this book, we have sought to ensure that each chapter can contribute to specific fields while demonstrating how those disciplinary contributions illuminate other domains. David Smith takes the lead in chapter three, for example, which subsequently includes more equations than a humanities reader might expect. However, that chapter also seeks to demonstrate how the computational methods it lays out for duplicate detection might pressure textual theories drawn from bibliography and book history. We hope that our readers will understand the multivocal disciplinarity of this book as a feature rather than a bug, as we hope to illustrate how a multidisciplinary project might advance both common and distinct aims for its constituent researchers. For that reason, we encourage all of our readers to explore all the chapters, even those that feel methodologically distant from their own training.

In this first chapter of Going the Rounds, we outlined the operations of the exchange system and argued that the reprinted texts in nineteenth-century newspapers are more than just filler, and instead comprise a wide range of everyday genres that evidence the themes and topics valued by readers during the period. What was “Going the Rounds,” in short, should matter to scholarly investigations of the nineteenth-century, and can expand our accounts or print culture to fragmented, partial, and amendable media largely unexamined by studied focused on books or even magazines. In chapter two, we describe a series of exploratory data analyses that examine the editorial and compositional practices of newspaper production through the nineteenth-century. This chapter seeks to demonstrate how textual data, gathered in large-scale digital newspaper archives, can help illuminate not just the content of historical media—our primary focus in most of this book’s chapters—but also its material features.

In chapter three, we focus on the material feature that has driven most of our work in the Viral Texts project, detailing our computational methods for detecting reprinted newspaper texts in digitized corpora. We propose a theory of “speculative bibliography” to describe our probabilistic, iterative approach to digitized historical materials. To capture the branching and chaotic operations of newspaper virality requires new methods: neither archival research nor keyword search are adequate to the task. The method of “speculative bibliography” we outline further in the next chapter is explicitly aimed to meet this need. Importantly, we do not posit speculative bibliography as a replacement for other methods. We would largely agree with scholars such as Matthew P. Brown that our method, which he would gather under the term “gross reading,” relies on inevitably flawed data sources, and that the broad patterns across many newspapers we identify necessarily obscure details about particular newspapers that would be gleaned from an “archival reading” approach to the same materials.36 The goal of a speculative bibliography, ultimately, is to direct scholars’ attention to textual configurations within mass historical data that might be worthy of closer study, an approach we argue models a productive conversation between distant and archival approaches.

In chapter four, we use our reprinting data to model the networks of exchange that connected nineteenth-century papers and their readers across time and space, showing that even authorship was subject to network effects, and was a function distributed across the exchange system rather than centralized in a few key figures. From here we use information from millions of detected reprints to model the exchange system as a network graph. We infer the shifting strength of connections between publications and locations from the texts shared between papers. Our networked representation of nineteenth-century US print culture echoes some familiar ideas, such as the prominence of Northeastern cities such as Boston, New York, and Philadelphia to the period’s print culture, but it also prompts us to take seriously the role of hub cities such as Wheeling, Nashville, or Columbus that have been overlooked by most book historical scholarship.

Our final two chapters home in on specific literary-historical questions. In chapter five, we investigate a genre we name “the vignette”: short prose articles that blend elements of reportage and fiction writing. We argue that vignettes constitute a proto-genre to literary journalism, and we employ computational genre classification to illustrate how the vignette draws from fact and fiction, and in what amounts. This chapter draws on the ambiguities of its computational methods to argue that genre itself is never a bright line, but an amalgamation of proportions and probabilities. Finally, chapter six draws on the Wright Collection of American Fiction to demonstrate how popular fiction was distributed and read through the exchange system.