Making the bi-directional argument of the dissertation required that I use a different medium than print. Books, and historical monographs in particular, operate according to genre conventions that encourage linear narrative development. Web documents, by contrast, operate according to different conventions — such as hyperlinking, interactivity, and layering — that create space for different kinds of argumentation. Because I am pursuing two inter-related arguments in this dissertation, and relying on data, code, and interactive visualizations to make those arguments, the conventions of web documents are a better fit for the work I am undertaking in this work.
For readability, the dissertation website does mirror some of the conventions of print dissertations. The main content of the dissertation is structured into four chapters, an introduction, conclusion, and the traditional supporting text of a bibliography and acknowledgments. However, although the chapters of this dissertation build sequentially, unlike a traditional narrative history, they are not structured to move forward through time or to unfold as a story. Rather, each chapter provides a different aspect of the overall analysis.
Additionally, the dissertation leverages the conventions of web documents to expose and connect the layers that make up the analysis of the dissertation. The site contains a collection of code notebooks that document the computational work behind the interpretive analysis as well as an interactive “browser” that provides access to the topic model at the heart of the dissertation. These components are linked to within the essays to expose the underlying computational analysis or can be explored independently for readers interested primarily in the methodological aspects.
As this dissertation relies on different conventions than traditional narrative texts, the interfaces require some explanation. The dissertation can be read in a number of ways. First, it can be read according to either of the two tracks identified on the website home page: “Ellen White and Gender in Seventh-day Adventism” for readers interested in the historical interpretation or “Computational Methods in History” for readers interested in the methodological arguments of the dissertation. Additionally, the project can be read according to the links in the top navigation, which guide the reader from the topic model browser, to the essays, code, and conclusion. Each of these elements is introduced below to explain how the pieces fit within the overall argument of the dissertation.
Topic Model Browser
At the core of the project is the topic model browser of the SDA literature, accessible at http://browser.dissertation.jeriwieringa.com. This model, built using MALLET and visualized using Andrew Goldstone’s DfR Browser, provides a window on the language used across the corpus of Seventh-day Adventist texts that I assembled. The interface is organized according to my interpretation of those topics, as reflected in the topic labels. I discuss my interpretive process for creating the labels in Chapter 3.
The four chapters that make up the primary text of the dissertation establish the context for and interpret the data from the topic model. These chapters build on one another and together offer the argument of the dissertation in a more traditional narrative form.
The first chapter, “The Emergence of Seventh-day Adventism,” provides an overview of Seventh-day Adventism in light of standard accounts of nineteenth-century religion and culture. This chapter situates the denomination historically, tracing their roots in the Second Great Awakening and William Miller’s predictions of the second coming as well as providing an overview of their distinctive theological and cultural features in relation to their development from a “sect” to a “denomination.”
The second chapter, “Constructing Computational Models from Historical Texts,” provides methodological arguments for the sourcing, processing, modeling, and visualization aspects of the project. This chapter is the most technical in content, tying the computational work to the historical interpretation.
The third chapter, “Anticipating the End of Time,” establishes the problem of time within early Seventh-day Adventism. A case study on using topic models to identify and reveal overarching trends in a corpus for historical analysis, I trace the prevalence of “end-times” topics over time. The study reveals three main cycles of end-times expectation that shaped the cultural development of the denomination.
The fourth and final chapter, “The Gendered Work of Salvation at the End of Time,” builds on the structure of time identified in chapter three to argue that the belief in the impending end of time created space for the development of a culture that was in tension with surrounding norms. The nearness of the second coming necessitated that men and women cooperate in the work of salvation, both within the home and in the missionary activities of the church, resulting in an expansive understanding of gender within the denomination.
As a project based on the computational analysis of text, a large portion of the intellectual and scholarly work that makes up the dissertation exists in code. In considering the methodological processes around the use of computational analysis in history, this dissertation makes the claim, both explicitly and implicitly, that this technical work is of central importance to the scholarly object of the dissertation. It does this through the inclusion of code and data files in a collection of notebooks, which document the computational work that went into creating and interpreting the topic model.
Where traditional methods of review rely on the use of clear and well formatted footnotes so that the trail of evidence can be retraced and evaluated, the review of computational work requires the use of documented and accessible code, so that both the execution of the techniques and the assumptions embedded within the technical choices can be seen and evaluated. The code elements can be viewed through the dissertation’s web interface or can be downloaded and run for verification or for adaptation to other projects. The work of making the technical elements of the dissertation visible and extensible is key to establishing processes for validating computationally-based scholarship in the humanities.
About: Process Statement and Bibliography
The final elements of the dissertation are the process statement and bibliography. Following the requirements of the Department of History and Art History at George Mason University, I provide an account of the methods and technologies used in the creation of the dissertation in the process statement, including the final websites, the modeling algorithms, and the dissertation sources. The bibliography lists the primary and secondary sources, as well as the computational tools used in the dissertation. This includes a full listing of the periodical issues used to create the topic model. Together, all of these aspects of the project constitute the body of work that is A Gospel of Health and Salvation.