In the third of this year’s Professionalization Spotlights, Alexander Clayton, the History Department’s public engagement and professionalization coordinator, connected with History PhD alum Jan DeWitt to discuss his career path and current work with the Massachusetts Institute of Technology’s InfoLab. In this transcribed discussion, Jan reflects on his work at U-M, the benefits of working in research, and what he brings to the table as a historian.

Alexander Clayton: Hi Jan. It’s great to get the chance to speak with you. To get us started, perhaps you could say a bit about yourself and your time at Michigan. When did you graduate? What did you research?

Jan DeWitt: I graduated in 2020 from the Interdepartmental Program in Ancient History, where I spent eight years working on the Roman Republic and civic development. My dissertation looked at the elected officials who actually ran the city of Rome and looked after the day-to-day functioning of an urban community. 

AC: What was your path after graduating? What led you to your current position at MIT? 

JD: I defended in January 2020, and by the end of May 2020 I was realizing that the pandemic was going to threaten many of the jobs in classics and history. It seemed like 80 percent of the jobs that I was interested in were just canceled in March and April. So I had to think about what I was going to do. I had friends in Boston who were able to connect me with Andrei Barbu at MIT InfoLab.

And so I dipped my toe into this neat new world of STEM research. The problem that Andrei and his team were facing was a medical one: when doctors are performing brain surgery on patients with epilepsy, it’s crucial to have a great neural map of where brain activity happens. The way they tend to do this is by having patients listen to audio and then mapping the different parts of the brain that light up.

But the problem is that children don’t have the attention span to listen for the amount of time they need in order to get a good neural map. Andrei and his team came up with the idea of using dialogue from movies. So my first role was that of lead annotator, where I spent the summer with a remote team taking movie dialogue that had been broken into four second chunks, and marking when every word began and ended, so that the team could get a clearer picture of what’s happening and where.

This past summer took us in a completely different direction. Having the largest collection of annotated film scripts has medical applications, but it also has uses for computer science.

One of the great difficulties in improving virtual assistants like Siri and Amazon Echo is teaching these computers to understand human speech. One of the powerful tools that we have for doing this is Universal Dependencies, which is a system for mapping out sentences. Most of the databases of sentences annotated using UD have been compiled using newspapers and the like. This is a problem though, because human beings don’t talk like newspaper articles, so the existing databases weren’t as useful as we’d like for teaching AIs how people communicate. Linguistically, movies offer interesting information because they present a more natural form of speech with interruptions, self-corrections, and incomplete sentences.

An example of a sentence from Spider-Man: Homecoming, mapped using universal dependencies.

Were there specific aspects of your PhD experience that encouraged you to look into this position in research and digital projects? Are there skills that have you been able to bring to the table as a historian?

JD: As a historian and a classicist, a really important skill has been my ability to learn new languages. My linguistics training was very useful for working out universal dependences and learning how to do sentence mapping. Whether it’s wading through a monograph or mapping an entire movie’s dialogue, having a good grasp of language and grammar has been a great skill.

More generally, there’s also a determination and adaptability that you gain from graduate school. Historians need to be determined and acclimate to working with difficult material. I found that I had a willingness to acquire the skills I needed, whether it was looking at CT scans and finding electrodes (this actually happened), reading a 300-page guide to universal dependencies, or now learning Python and coding.

I could never have expected where these last two years were going to lead, but I have enjoyed the fact that there is a very tangible benefit to what I’m doing. In the first part of the project, helping people with epilepsy. In this latest part of the project, improving apps that we use every day. In the context of the pandemic and all kinds of turmoil that we’ve been living through, that has been really rewarding.

AC: Have there been any learning curves that you’ve had to overcome? What have you learned and gained from working with such an interdisciplinary group? How has this altered your perspective as a historian?

JD: The main lesson is that there is a lot of really interesting work to be done that is not publishing articles or monographs. These tasks have been really challenging, fun, and rewarding. You don’t necessarily have to be your own researcher who is publishing, there’s lots of fulfilling work out there. 

If there’s one thing I could have done differently, I think I would have started acquiring these skills sooner. There are definitely ways to use digital methods and coding to better understand language, personal networks, and geographies. Computer models absolutely would have been useful in the prosopographical work that went with my dissertation.

AC: Do you have any advice for current graduate students that are considering a fuller range of  careers outside of the tenure track? Are there ways that you would recommend exploring new paths, making connections, and gaining experiences?

JD: Especially with the job market as it is, it’s important to not have a blinkered approach. Don’t limit yourself to your current skill set, and look for ways of expanding your research and methods to gain broader skills. Coding is a great example of that, it’s something you can use for digital humanities and in the professoriate, or in a wide range of careers.

And as much as I hate to be the person who plugs networking, it’s definitely the way to go. I’ve honestly found that the best networking is done when you don’t think of it as a network, just stay in touch with friends and colleagues and check in with people. It’s not only a great way to learn about opportunities, but also to stay happy.