For our Boot Camp, we worked together with programmers to create a database of the epistolary texts present in Bram Stoker’s Dracula.

Alanna on her conception of the Dracula database project:

This project was a bit like diving into the deep end of a pool, never having seen water before. This is my first foray into the digital humanities, and it raised a lot of interesting questions over the course of the project. I considered whether using a digital tool for a Victorian project necessarily made it part-digital humanities. Does using digital tools make a person a digital humanist? Is a child using paints and brushes less a painter than Van Gogh?

Methodology

I decided to approach the project using a rough approximation of the scientific method. Just thinking about using a digital component changed my methodology—I was no longer only wearing my Victorian literature hat, I was heading into unknown territory and needed a helmet. For this reason, and also the sneaking suspicion that I had no idea how much I didn’t know, it seemed smart to document the journey as a sequence of hypotheses, predictions, experiments, and conclusions.

The Question

To build a digital tool for literary criticism is an excellent thing to do if it can contribute meaningfully to one’s research. In other words, does the tool have a function? It took some time to come up with the questions I wanted the tool to help me answer. I also reflected on why I wanted it – the digital humanities are a fashionable field, and it is easy to be drawn to them because there are lots of beautiful and interesting innovations, but ultimately, to build a database that does not serve a specific purpose is to build a toy, not a tool. 

This project emerged from a question that came to me during the research for my MRP, Dracula’s Private Collection. What might happen when an epistolary novel, in this case Dracula, is restructured or reorganized in different ways? To find out, we built a database that would sort the items in the text by date, author, and genre, to see how the narrative might change. How might the text change in meaning if one were to read only the diary entries or letters? Does the text change in any significant way if it is reorganized to be chronological, or if we only read the men’s writing? 

Beyond that, there is also a larger question concerning the archival nature of the novel. The novel is a collection of letters, diary entries, journals, newspaper clippings, and various other items that Mina Harker compiles into the archive and is ultimately the weapon they use to defeat Dracula. Dracula himself is something of an archive as well, because he is a masterful collector of languages, victims, and blood. Mina’s modern archive (modern because it is type-written by a woman) is used to slay this ancient, occult one, and so questions of archival violence begin to arise. Derrida’s concept from Archive Fever is useful here because authority is key.

Hypothesis: By creating a tool that can change the internal structure of Dracula, we will in turn create shifts in meaning that will yield new insights into the text.

Prediction: That this database will not only enable deeper study of the text, but also open up new ways of examining its structure and the metafictional constructs that make up the novel. 

Experiment: Create a database, and use it to perform analyses of various sets of data.  

Hilary and Jess on their experience working with the database:

In her probe for this week, Hilary wrote about the specific materiality of objects in the archive or “database” of a library, museum, or one’s personal collection—the material details that set them apart from one another. Just like any epistolary novel, Dracula’s objects of correspondence are catalogued within the pages of the book, made alike via their appearance on the page (the same font and spacing, etc.), but we, the reader, know that they are not alike in spirit. A letter from Mina or a telegram from Harker will feel like very different entities in our mind’s eye, as we read the story, because we extrapolate from what Stoker gives us into the signification of a letter, or a telegram, in all its ephemeral materiality. Yet the words of these documents also produce in us an affective response that sets them apart from one another. A telegram connotes urgency and is less intimate than a hand written letter, onto which we may imagine wax spilling and congealing or fingers smudging. Is consideration of such materialities (or lack thereof) considered distant or close reading?

To put this in conversation with the Manovich article, we can ask ourselves whether we are really de-materializing Dracula. Arguably, we are materializing it. To put it even more generally, we could question what our role is here, versus that of the programmers. (Especially given that the gender divide between us and them only served to further emphasize the gap in our knowledge). Here, for your (sadly?) comedic viewing pleasure, a photo that illustrates the chasm between us:

  Are we interpreters of this data working with an entirely different paradigm-syntagm relationship? Haven’t we experienced a bit of an inability to communicate between disciplines because of this difference?

With these distinctions in mind, we approached the Dracula database project eagerly, hoping to learn something about code and, even better, to force ourselves to be interested in such an undertaking. While we did find the process fascinating, it was less because we became genius code masters, and more due to our interests in the mistranslations between us, as literature students, and the programmers. At one point, the three of us were talking about the “violence” that this type of project does to a work of literature like Dracula. The fact that these intimate epistolary exchanges between Stoker’s characters be reduced to a series of “stringlines” does seem violent, in a sense. Richard, the programmer, took issue with this word “violence” that was being used to describe his pragmatic pulling apart of the text, a reaction that interrogated our academic impulse to use such loaded words. The act of configuring a work of literature into a database in particular, is violent in that all the signified values conveyed by the language to the reader are eradicated and made useless. You can’t search for a document in the Dracula database by their affective or signified qualities (ie. “letters tinged with jealousy”)—it is only possible to identify entries by the simple imagistic qualities of the words that are present. A date, for example, has the date-word, the month-word and the year-word. There is nothing special about “October” other than the fact that it has an O followed by a C and etc. Lev Manovich addresses this when he writes that new media objects appear “as a collection of items on which the user can perform various operations: view, navigate, search. The user experience of such computerized collections is therefore quite distinct from reading a narrative or watching a film or navigating an architectural site” and that the result of the database is a “collection not a story” (2, 3). Manovich writes that “computer programming encapsulates the world according to its own logic” and he distinguishes between two types of “software objects” which this world is reduced to: “data structures and algorithms” (6).

One of the first things Richard the programmer showed us was the file of Dracula that he had found on Project Gutenberg:

Questions of provenance, text form and text edition came up in discussion of which version of Dracula was being fed/scanned into the database. Should we question the authority of a project like Gutenberg on such matters, we wondered. Does it matter that we’re working with a digital text rather than manually scanning/transcribing a physical copy? Is there a different materiality at play? Even with the physical book there is already a sense of duplication—the narrative is a sum of textual copies; Dracula is also a creature seeking replication, and as immortal and theoretically vanquishable as a digital copy. Jess mused that she found it fascinating to think of this database project as just another vehicle for Dracula’s world contamination and pathology in that it gives new meaning and agency to his monstrous textuality, which we are in effect enacting. In a weird way, in decomposing the text through the archival process, the text itself is becoming more real—an uncontrollable thing that both necessitates human activity to bring the database to life, and yet in its potentially endless unfurling/mutation becomes increasingly inhuman as the text-objects are dissociated from narrative patterns.

In order to “read” the markup language attached to file, the programmers needed to try different reader programs such as Eclipse, Java, then Python, and finally settling on C sharp.

They used an addition to C Sharp called Re-sharper which improves the function of the program. Richard told us he would have to delete the table of contents from the file’s markup language as it was superfluous to the database project (this is one place we talked about “violence”). Richard briefly explained object-oriented languages to us—first you create classes, such as “file reader,” and then objects, which must have the same name as a class. Once you’ve created your attributes for said objects, you can examine the strings of the file (or the sequence of characters, such as a sentence) and look for patterns which may be useful to you. Then you must create conditions, using words like “while” and “else” that will allow you to act upon the objects. Because we wanted the database to separate the entries by name, date and medium, we needed to create conditions that would act upon the stringlines within the file and pluck out all Journal Entries, for example.

The Future of this Database

HTML v. XMLTEI

The text is HTML, which is ultimately a code for presentation, for visual reference. The consequence of coding for display, rather than coding for meaning, is that this HTML version is really only useful to Alanna. It does not contribute to the larger scholarship, and it does not allow for other scholars who wish to manipulate Dracula digitally to do so except in very narrow ways. It might be fun for undergraduates who are working on Dracula but for larger projects, its usefulness is limited. The way to correct this is to create a version using XMLTEI because that is a language that codes for meaning. It looks almost exactly like HTML, and if you know HTML you can learn TEI quite quickly, but the reason it is so much more useful is because you can code for anything. If you want to create a personography, you can create TEI PersName tags around each occurrence of a given character’s name: “One day, <persname> Alanna <persname> walked down the street.” This allows for navigation through the text in a way that follows one storyline, because now that name has been linked semantically to an entry in a personography. Anything can be coded: names, dates, characters, objects, particular words, even smudges and marks on the paper of a given manuscript.

There will inevitably be more coded values and meaning than what any given scholar might use. The value of this is that the tags are never taken out, and that means that even if one person decides that dates are not important, someone else who does can do so with relative ease.

The way to use XMLTEI is to first check if the text you want to work on is an XML document that is also TEI compliant, or if it is just HTML. This information will usually appear in the header. It is also necessary to learn XSLT, which is a way of navigating TEI. Software developers will laugh because these are “so 1993,” as one recently put it. However, TEI has evolved into something of a specific digital humanities language, and so what is obsolete in the software world is not in the digital humanities. It is the most used language before making the jump to something like MySequel (a database language), and Python or one of the other high-performance languages.

To use these tools on Dracula and on other Victorian gothic epistolary novels will take a significant amount of time and also probably money, as well as training in TEI and XSLT. For these reasons, it is not practical to undertake them now for the sake of upgrading the Dracula database, but it does open up various possibilities for a larger, more sustained project in the future.

This brings us to yet another line of query: Is deconstructing Dracula in this fashion helping us experience the phenomenological horror that is Dracula on another level? Are we seeing the text assuming a new autonomy? Accessing Dracula through such a database in a way means breaking away from reading habits such that we are confronted with the impersonality of all texts. There is a disturbing interaction here between human and nonhuman elements that the database-building process reveals.

How much does the human recede into the nonhuman or vice versa?

Are we not seeing our own scholarly desires being translated into code?

Are we not probing into language’s otherness and translating it?

Some of the words that the programmers used interested us, such as “ugly” to connote variance (for example letters on which the dates deviated from the standard format, “day, month, year”). Whereas for Richard, “violence” was a hard word to swallow, Hilary noted that for her, “ugly” felt the same. Perhaps “deviant” would be a more accurate term? But that too implies that there should be some sort of standard formatting for dates of handwritten letters—an assumption that further erases the individual, human (or, in this case, character), handwriter from the letter-writing process (and the materiality that we continue to cloyingly hearken back to). The names of the programming languages also interested us. One of them, called “Lisp,” uses brackets for comments and apparently isn’t known for running smoothly. The name amused us as the implication of speech impediment made clearer the distinction between the human user of these languages and the coded language itself—as well as the awkwardness that may ensue when the two attempt to grapple with one another.

Jess wondered whether there wasn’t a kind of terrifying repetition to this whole process— or a sense of never really moving forward, and how this might offer a different experience of textuality? Dracula is itself a terribly repetitive narrative. We wondered whether, by participating in this project we’re looking for something external to the text— not part of the text itself but peripheral to it that the mechanics of archiving would render visible?

Graham Harman’s article, “The Well-Wrought Broken Hammer: Object-Oriented Literary Criticism,” in which he addresses the autonomy of textual objects from their relations and New Criticism’s holistic conception of the text’s interior, can speak directly to our project. Harman questions which features of a text or work of literature are essential to its definitive objecthood:

We can add alternate spellings or even misspellings to scattered words earlier in the text, without changing the feeling of the climax. We can change punctuation slightly, and even change the exact words of a certain number of lines before [the text] begins to take on different overtones…the critic might try to show how each text resists internal holism by attempting various modifications of these texts and seeing what happens. Instead of just writing about Moby-Dick, why not try shortening it to various degrees in order to discover the point at which it ceases to sound like Moby-Dick? Why not imagine it lengthened even further, or told by a third-person narrator rather than by Ishmael, or involving a cruise in the opposite direction around the globe?…In contrast to the endless recent exhortations to “Contextualize, contextualize, contextualize!” all the preceding suggestions involve ways of decontextualizing works, whether through examining how they absorb and resist their conditions of production, or by showing that they are to some extent autonomous even from their own properties. Moby-Dick differs from its own exact length and its own modifiable plot details…there is a certain je ne sais quoi or substance able to survive certain modifications and not others. (Harman 201)

Similar questions can be asked of our project—which uses such an object-oriented database model: are we stripping away at all the inessential qualities of Dracula? Are we left with anything that is essentially Dracula? That makes it recognizably such? Having said earlier that we are expanding the text of Dracula or metastructuralizing it it’s weird to think we are in the same gesture stripping it down to its bare parts.

By now, it has probably become clear that our boot camp project raised many questions for us, but these are, in fact, the types of questions that only a project such as this would fully force us to ask. In the face of the inevitable “So what?” we could respond that we have attempted to take Bogost’s suggestion to heart—that to build something can help you better understand it (or, in our case, help you realize how little you understand it)—that a project which represents “practice as theory” (111) is a worthy one, regardless of the outcome.

Works Cited

Brazeau, Bryan. Personal interview. October 3 2014.