Charlotte's Web: Encoding the Literary History of the Sentimental Novel

Charlotte’s Web: Encoding the Literary History of the Sentimental Novel
Melson, John, Women Writers Project, Brown University, John_Melson@brown.edu
Funchion, John, University of Miami, jfunchion@mail.as.miami.edu
Ever since the cultural turn in literary studies, literary scholarship has focused on examining the cultural work performed by texts. Indeed, one of the enduring theoretical innovations of the 1980s was the reconceptualization of literary value according to the concept of cultural work—the idea that texts should be evaluated not by the innate aesthetic or formal qualities they manifested, but according to their “designs upon their audience” and their ability “to make people think and act in a particular way” (Tompkins 1985). As a result, previously overlooked or marginalized texts have often been deemed suitable for study in the ensuing years based largely the on degree to which they are said to perform particular kinds of cultural work. In one familiar example from the arena of nineteenth-century American literature, scholars have come to see a stereotypically sentimental novel like Uncle Tom’s Cabin as a canonical work primarily because its complex entanglements with the politics of the antebellum United States allow it to be read— to cite Tompkins once more—as part of a “monumental effort to reorganize culture.” But while this approach has become an important characteristic of much literary scholarship, the question of how to measure a given text’s capacity to perform cultural work remains vague and often myopically focused on the synchronic significance of the work in question at a specific point in time.

In this paper we investigate the matter of cultural work from the vantage point offered by current scholarship in the digital humanities. We do so with an eye toward developing a model of literary history that draws on uniquely digital methods for structuring and formalizing intertextual relationships, and that make it possible to chart the cultural and formal significance of literary works across space and time. Specifically, we use Susanna Rowson’s late eighteenth-century sentimental novel Charlotte Temple as a case study in how digital literary history can evaluate the meaning and formal significance of a text diachronically. Often cited as one of the earliest American bestsellers, Rowson’s work was reprinted hundreds of times during the nineteenth century and was referenced repeatedly in an extensive body of writing, ranging from reviews and advertisements to melodramatic stage adaptations and ostensibly factual regional histories of the United States. Tracking the novel’s extended nineteenth-century afterlife through this complex network of external references, we demonstrate how the application of detailed interpretive markup to the multiple documents that reference Charlotte Temple (instead of the novel that would, more conventionally, be thought of as the “primary” text) enacts a theory of literary history in which concepts of cultural work may best be observed as phenomena that develop over time.

While the idea of interpretive markup is not new, it is most often used either as a means of recording a layer of scholarly annotation on some particular document or classifying portions of a document according to some external taxonomy. Both uses are supported, for instance, by the current version of the Text Encoding Initiative (TEI) Guidelines, which provides examples of specific methods for encoding “semantic or syntactic interpretation” using a set of standard TEI XML elements and attributes. In practice, though, discussions of interpretive markup in digital humanities projects often revolve around questions of readability and the appropriateness of stand-off versus inline markup (e.g. does excessive interpretive markup pose problems for human readability of encoded text? [Campbell 2002]), or issues of preservation, curation, and reliability (e.g. what challenges for encoding textual information accurately and reliably are created by allowing extensive interpretive markup? [Berrie et al. 2006]). Although such questions are important in certain contexts, our project treats them as less important than the question of how interpretive markup may be enlisted as a surface for scholarly analysis—that is, how interpretive structures can themselves be classified and interpreted.

Our project adapts several of the interpretive structures provided in TEI XML to represent the connection between specific formal properties in Charlotte Temple and external evidence of the novel’s broader cultural influence, as attested by external references to it. We record basic metadata about each external reference (author, title, date and location of publication, etc.) while also indicating which specific features of the novel the reference comments on, as well as the nature of that commentary. The result is a set of XML-encoded documents classified according to the interpretive work they do: a record, as it were, of how nineteenth-century readers interpreted and responded to the novel’s formal properties. At the same time, a secondary layer of interpretive markup further formalizes the relationships we identify across this primary layer of interpretation. Taken together, both layers provide substantial material for further abstraction: for instance, the automated generation of topic maps that represent a variety of cultural knowledge structures. The result is a web of connections in the form of citations, references, allusions, parodies, and comments spanning the nineteenth century, whose formalized structures constitute the interpretive tissue of digital literary history. The ability to map and visualize these structures, we argue, offers significantly new possibilities for observing cultural work as a phenomenon that evolves over time, and whose connection to particular texts is most productively understood as operating at the nexus of close reading and “distant reading” strategies.

Our work constitutes an initial small-scale experiment, but it has relevance for larger questions of scale, purpose, and method in both the digital humanities and conventional modes of literary scholarship. In particular, it suggests that although digital scholarship offers a potential reconceptualization of literary history, practices of literary history also pose interesting challenges for work in the digital humanities. In recent years, projects in the digital humanities have increasingly responded to the question of scale by refining methods for aggregating, classifying, and analyzing enormous bodies of textual material—in other words, by treating text as a mass of data that can yield meaningful responses to statistical analysis of its language. Whether offering answers to now-familiar questions about scale itself— “What do you do with a million books?”—or taking up Franco Moretti’s challenge to peer into the “cellars of culture” represented by the tens of thousands of unread and unknown novels published in the nineteenth century, text mining and other forms of “computational humanities” have been increasingly held up as a means of “checking the generalizations of literary history”(Crane 2006; Parry 2010; Pasanek and Sculley 2008). At the same time, long-running digital humanities projects with deep investments in scholarly text encoding have tended to approach literary history from the opposite direction, emphasizing how “craft encoding,” for instance, employs editorial methodologies and modes of textual scholarship that participate in “making literary history” in the digital medium (Flanders 2009; Dimock 2007). While neither approach directly contradicts the other—indeed, in practice they often coexist harmoniously—they inflect the concept of literary history in crucially different ways. The former often treats the machine-processable language of texts as a primary axis along which historical change in writing manifests itself: intertextual relationships are revealed in changing patterns of linguistic borrowing, with the emphasis on facilitating the “comparability of texts” (Mueller 2008). The latter, while still valuing intertextual comparison, tends to prioritize the documentary and contextual over the linguistic—for instance recording textual variation across editions of the same text, or formalizing through encoding “rich environmental contextualizations” for the study of particular texts (Folsom 2007). Our project borrows from both strategies, and in doing so seeks to demonstrate in practice a method by which digital literary history negotiates this apparent bifurcation.

References:
Berrie, Phill et al. 2006 “Electronic Textual Editing: Authenticating Electronic Editions, ” Electronic Textual Editing, Burnard, O’Keefe, and UnsworthNew York: MLA (link) .

Crane, Gregory2006 “What Do You Do with a Million Books?, ” D­Lib Magazine, 12 (3)[ (link) ]

Dimock, Wai Chee2007 “Introduction: Genres as Fields of Knowledge, ” PMLA, 122 (5)1377-88

Flanders, Julia2009Seminars in Scholarly Text Encoding with TEI. Paper presented at the 2009 TEI annual conference, , Ann Arbor, Michigan. (link)

Folsom, Ed 2007 “Database as Genre: The Epic Transformation of Archives, ” PMLA, 122 (5) 1571-9.

Mueller, Martin2008 “TEI-Analytics and the MONK Project, ” Paper presented at the 2008 TEI annual conference, , London, UK. (link)

Parry, Marc28 May 2010 “The Humanities Go Google, ” The Chronicle of Higher Education, . [ (link) ]

Pasanek, B. and D. Sculley2008 “Mining Millions of Metaphors, ” Literary and Linguistic Computing, , 23 (3): 345-60.

Text Encoding Initiative. P5: Guidelines for Electronic Text Encoding and Interchange, (link)

Tompkins, Jane1985 Sensational Designs: The Cultural Work of American Fiction, 1790-­1860, New York: Oxford University Press

Wilkens, Matthew2009 “Corpus Analysis and Literary History, ” Paper presented at Digital Humanities 2009, , College Park, MD.