We would like to extend a warm welcome to our new Digital History Fellow Jana Keck. Her fields of research include German and American Literature and Culture in the Nineteenth Century, Periodical Studies, and Digital Humanities. She began her academic career at the University of Stuttgart, where she studied English and Linguistics. In 2017 she received her Master’s from the University of Stuttgart with her thesis “Gottfried Duden’s Bericht über eine Reise nach den westlichen Staaten Nordamerikas (1829) and Its Emigration Stimulus.” In the same year she joined the DFG-funded research project “Oceanic Exchanges: Tracing Global Information Networks in Historical Newspaper Repositories, 1840-1914” as a Doctoral Researcher. She is also a member of the CRETA/Center for Reflected Text Analytics, funded by the German Ministry for Education and Research (BMBF).

Recently we sat down with Jana to discuss her academic career, her work, the field of digital history and its impact on the field of humanities, as well as her current goals and duties at the GHI Washington.

Could you give us some insight into the Oceanic Exchanges Project?

In 2017 I started working with an international team in the DFG-funded research project “Oceanic Exchanges: Tracing Global Information Networks in Historical Newspaper Repositories, 1840-1914.” The project was striving to gather researchers not just from different academic fields such as cultural historians, computational linguists, literary scholars, digital curators and humanists, but also from seven different countries. Together we collected over 100 million pages of digitized newspapers in order to examine how, in the nineteenth century, news, texts, and concepts traveled, how stories adapted in different environments when they traveled, and how news networks functioned.

How and why did you become interested in Digital History?

I became interested in the subject through my PhD supervisor when I was studying English and American Studies and Linguistics at the University of Stuttgart. While working on my Master’s thesis on German travel literature to the United States in the nineteenth century, my PhD supervisor told me about this upcoming project in digital humanities and wanted me to be part of it. However, since I had not studied digital humanities, I knew there was a lot of “self-study” and “catching up” to do.  So, I took some courses and, in doing so, I really became interested in programming. I think it was because there was a familiarity to it that reminded me of linguistics, especially the focused and logical aspects that make up linguistics. I realized that it gave me the opportunity to combine my different fields of interest and not be relegated to one field. The more I went out and saw how other fields were working, the more I gained a deeper understanding of my own field. That is how it all started. I attended programming courses, workshops in computational linguistics, and I started working not only with people from the humanities but also from other fields not traditionally associated with the humanities. What I love about it is not only that it is a “newer” discipline, but also that I get to work with people from different fields and that I am able to take on different roles. We often define the humanities as one big field, but it has as many differences as History and Computer Science. Historians may define a certain term one way while linguists use it completely differently. I find it important to communicate these differences to make what we are talking about more transparent.

Can you tell us a little more about your PhD project?

I was inspired to begin my project by the Oceanic Exchanges Project but also by one of the former Projects called “The Viral Texts Project” initiated by Ryan Cordell at Northeastern University in Boston. Cordell was studying reprinted texts in nineteenth-century American newspapers, which prompted me to focus on nineteenth-century German-language newspapers and the phenomenon of editors sharing texts. One of the most interesting questions was whether this phenomenon is to be found among migrant newspapers and, if so, what kind of texts they shared. This means that I am not interested in one particular newspaper but rather in what newspapers printed in German in the nineteenth century in the U.S. have in common. My hypothesis is that German newspapers were used as a form of community building, since there was not one German identity among immigrants from German-speaking Europe. Germany was then only a geographical expression. The immigrants would have defined themselves in more local or regional terms: such as from Baden or Bavaria. Once they arrived in the US, however, they were seen as monolithically German by the receiving culture. What they had in common as opposed to their “American” neighbors was their “Germanness.” Through the newspapers, they developed a shared German migrant experience on both the local, national and even transnational level. Analyzing the information they exchanged through the newspapers – from hard news to jokes, from emigrant lists to poems – sheds light on how they constructed a German community across states and decades. It reminds me a lot like social media platforms: so to speak, I am investigating how a specific community shares their tweets and how they change once they travel through time and space.

What DH methods do you apply and what are your findings so far?

I began with a corpus of 60 newspapers from 17 different states over a period of roughly 80 years. I am using text reuse detection algorithms to identify and cluster texts that were printed several times in the same and in different newspapers. However, the amount of data available made it impossible for me to read all the material closely, so I found a way to computationally classify these different texts into different genres. This genre classification will make it possible for other researchers in the field to access and work with the material in the future. Currently, I am working on combining different supervised and unsupervised methods to classify different texts on the basis of the training data set, for example, indicating whether they are poems or hard news, etc. The next step is to determine if the system is also able to classify these data sets if they are previously unknown to the program. It is working very well in differentiating between news versus fiction and even within the news section between lists and advertisements, for instance. However, it runs into problems when it comes to genres in fiction. It is difficult for the system – but not impossible – to differentiate between a poem and a joke. I am positive that I will find a way to fix this and if not, it is one more finding on a theoretical level.

What other Digital History projects are you involved with at the GHI?

I am still new to the institute, but I already have been involved in a lot of meetings regarding the German Heritage in Letters Project. One of the visions I have is to combine the newspapers and the letters in order to find out what they have in common: from people to ideas. Also, an incoming DH fellow wishes to use the fellowship provided by the GHI and the Roy Rosenzweig Center for History and New Media and funded by the Gerda Henkel Foundation to create a database of German immigrants in the same timeframe. It would be very exciting if we could combine the three different projects that all fit so well into the current research profile of the GHI which is focused on Digital History, the History of Knowledge, and the History of Migration and Mobility. To work in such a stimulating context is just very rewarding!

Interview by Alexa Lässig, with thanks to Sarah Beringer and Casey Sutcliffe for the introduction and proofreading.