/etc

tags: talks
Originally Published: 2015-03-26

Below is the text for my Shifting Frontiers in Late Antiquity XI keynote speech “Shifting Frontiers in the Digital Humanities”, delivered at the University of Iowa.

The accompanying slides are available here.


I’m honored to be here talking to you today, kicking off Shifting Frontiers in Late Antiquity with a “Digital Humanities Keynote”. Particularly honored as an opening keynote gives you the opportunity to weigh in with your opinions before so many intelligent, brilliant scholars present theirs. So, there’s very little risk of boring you by repeating something that’s already been said.

Now, you may object, that might be true if my topic were something specific to late antiquity, but the title of my talk has “the digital humanities” in it, and what hasn’t already been said about “the digital humanities”? With so much ink (and data) spilled, what remains to be said, what position remains to be taken? In trying to stake out new territory, you would have to survey the the existing landscape. There, to my eyes, lies a vast and tangled jungle of interrelated topics:

What are the digital humanities? What are the humanities? What are the humanities good for? What are the digital humanities good for?

Now, I would argue that none of these are particularly new or particularly unique to the humanities. Questions of meaning and purpose have vexed people for ages (just ask philosophy), and ontological questions about the bounds that draw a discipline have confounded people as well (just ask philosophy of science). I don’t think I have the time here to settle any such question, and as tempting as it is to claim that it’s simply because “I have a marvelous proof which the margin is too small to contain,” the truth is I would be skeptical of anyone who claimed they had definitively and simply resolved these problems.

Instead, I want to take an optimistic, pragmatic look at the sorts of opportunities, changes, and opportunities for change that people wrestling with these problems have produced. Just as the call for papers emphasized the parallels between modern and ancient shifts in society, I want to highlight how these shifts are taking place in the “digital humanities” as well.

I think “shifting” is also an apt term to use in these contexts. It avoids the sort of “Whig history” implied by “progress”. Progress is relative, hard-fought, and sometimes lost.

The first example of a sort of progress or shifting I’d like to start with is one from my own work, and one which probably relates at least somewhat directly to the core topic of this conference. I’m part of the Duke Collaboratory for Classics Computing (or DC3 as we like to call ourselves), and one of our main projects, in fact probably the reason for our existence, is a website called papyri.info. It provides online editing and search for papyrological documents. These are primarily documentary papyri, so they’re fascinating texts that give us a glimpse into day-to-day life in the ancient world - letters, tax receipts, legal documents, memos about grain…okay, maybe not as fascinating as they could be, but these are the things that make up day-to-day life. We have, according to my count just now, about 60,000 text transcriptions encoded in EpiDoc XML - about 30,000 of these, according to our metadata, are potentially datable to the period between 200 and 700 CE.

To understand the shifts in this project, you first have to know its history. The primary corpus of transcriptions is from a digital project that started in 1982 called the Duke Databank of Documentary Papyri, or DDbDP. You heard that right, a digital project that started in 1982. In 1996 the project migrated from betacode to SGML. From 2007-2012 a Mellon-funded project called Integrating Digital Papyrology or “IDP” took place in three phases, to support transitioning from SGML to EpiDoc XML, map relationships between papyrological databases, enhance search, and build an online collaborative editing environment. The vision of this, along with new editorial infrastructure, was to “allow the entire community of papyrologists to take editorial control of core disciplinary data.”

However, another goal with this, or perhaps another way of accomplishing it, was to not just provide papyrologists with an online interface by which they could search and contribute to a closed data silo. The goal was to also enable true community ownership of the data. So while this is a nice interface that provides an easy way for anyone who wants to to enter and edit papyrological texts, and it took a lot of time to build, and takes a lot of time to maintain, it’s not really fundamentally itself that revolutionary. What I’m excited about is that this is just a front-end to a dataset which is completely open and available under an open license. So anyone can download the entire dataset and do whatever they want with it. More than that, they can see the complete revision history, continue to pull in our updates if they make their own changes, and even easily submit their changes back to us if they like.

What we can see here from the revision history is a line-by-line analysis of who last modified each line in a text. I like this because it very clearly illustrates the problem of assigning a single “editor” to an edition of the text. The text as constituted at any point in time is the result of a complex dialogue. So being able to easily represent and analyze this, is a shift for scholarly work. As my colleague Hugh Cayless likes to say: complexity is irreducible—it can be moved, even hidden, but it must be faced. The shift here, for us, is a scholarly decision not to hide this complexity.

Another shift for the Duke Databank in this is to be explicit about the editorial processes involved. The papyrological editor attempts to digitally replicate a system of scholarly peer review, so that the data published under our imprimatur has been vetted and approved, along with an attributable record of the process. But again, nothing stops anyone from using, modifying, and republishing our data however they want. We really do want community ownership of this data.

All of this, including the ways we implemented editorial review, were the result of hard deliberation. Part of this difficulty is because making deterministic, algorithmic rules for what were previously social, human processes can be inherently hard. A common aphorism from engineering is that there are rarely easy technical solutions to social problems. Human dialogue and social interaction can go a long way towards resolving issues that it’s hard to design an algorithm for. But I also see this reflected through the very fabric of our project. We’ve designed it in such a way that the social convention of the authority of “publication” is explicit. We don’t try to set up an artificial barrier in order to establish and retain our authority over what we see as fundamentally community-owned data. If these texts are our shared cultural heritage, aren’t they, by definition, owned by all of humanity?

I think it reflects poorly on scholarly publishing that this view is, in fact, a shift from the previous status quo. As tempting as it is to say that this was merely an artifact of the costs of physical print media, examining the constitution of various digital scholarly publishing projects will prove otherwise. This shift in view is not a given, it’s hard fought, and tied up with the never-ending dialogue of what we, as a society, want to value.

As an illustration of this, and to shift from how the present studies the past to how the present studies the present, I’d like to show something which I think is exemplary of the digital humanities reflexively addressing modern concerns in modern society. This is Ben Schmidt’s “Gendered Language in Teacher Reviews” project, which aggregates data from 14 million reviews from RateMyProfessor.com. What this enables is a very fast, easy visualization of gendered, sexist language and biases in aggregate from a large data set. So here we can see what are obviously two heavily gendered negative terms, as well as one where everyone fares about equally. One interesting facet of this is that it quantifies information from what are ostensibly qualitative review texts. This, I think, is important: if these same students are writing the student evaluations which are used by administrators for decisions about tenure, promotion, and retention, then gender biases in these evaluations, even if they’re unintentional or unconscious, may perpetuate a sexist culture. So knowing and understanding about these biases is important, if society wants to address these problems.

There are also shifts occurring in the digital humanities itself, addressing concerns about anglo-centrism, the sorts of communities we want to foster, and the biases that we might unconsciously hold, ourselves. Postcolonial Digital Humanities, or DHPoco, is co-directed by Roopika Risam and Adeline Koh to explore some of these threads. Global Outlook::Digital Humanities (GO::DH) is a special interest group of the Alliance of Digital Humanities Organizations, whose stated purpose “is to help break down barriers that hinder communication and collaboration among researchers and students of the Digital Arts, Humanities, and Cultural Heritage sectors in high, mid, and low income economies.”

Now, as perhaps a more literal illustration of shifting frontiers in the digital humanities, and returning to digital resources which directly address late antiquity, I want to highlight some recent developments in geospatial resources for the ancient world. The Pleiades project is an exemplary model for digital humanities projects. Started in 2006 with support from the NEH, Pleiades provides a digital, community-built gazetteer of ancient places from the Greek and Roman world, and it aims to expand its coverage “into Ancient Near Eastern, Byzantine, Celtic, and Early Medieval geography.” But this Western classical tradition is not the only one, and resources to address this gap are starting to arise alongside increasing acknowledgment of the influence and value of other traditions and cultures. Here you can see an early version of a project called Al-Thurayya by Maxim Romanov at Tufts, which aims to extend the ancient place gazetteer concept to the Islamic world. There’s also a project called the Syriac Gazetteer at syriaca.org run by David A. Michelson, for “places relevant to Syriac Studies.” If you want to discuss shifting frontiers in late antiquity, both literal and societal, this seems like a fairly relevant triumvirate of geospatial resources for the Mediterranean and Middle East. But the existence of these resources was not a given: Pleiades staked an initial, experimental claim, in the hopes that such a resource would be useful to others interested in the topic, and other resources are now forming in an amiable dialectical reaction to and expansion upon this resource.

So, I said I wasn’t going to answer the question “what are the digital humanities?”. I lied. The digital humanities are what people, including you and me, make them. I hope I’ve highlighted some ways that that’s happening, and inspired you not to let any received rhetoric of what is or isn’t possible limit your imagination of what’s possible. The digital humanities are a shifting frontier. People are shifting. Society is shifting. Humanity is shifting. So, the humanities are shifting as well. Just the way they always have.