symbolic image newspapers

Facebook in the 19th Century

forschung leben – The Magazine of the University of Stuttgart

Digital methods are used at the University of Stuttgart to research the distribution of information through newspapers in the 19th century. Computer-aided research facilitate to investigate patterns of information flow across national and linguistic borders.
[Photo: Jana Keck/ILW/University of Stuttgart]

An international team is analyzing how newspaper stories used to spread, in the “Oceanic Exchanges” project. University of Stuttgart researchers from the fields of literature, computational linguistics and computer science are participating in the project.

When Krakatoa, a volcano between the islands of Sumatra and Java, erupted on the 27th of August 1883, people in Europe and other continents found out about it the very next day: they read about it in the newspapers. “That eruption is considered the first global media event”, says Prof. Marc Priewe, head of the Department of American Literature and Culture at the University of Stuttgart’s Institute of Literary Studies (ILW). How could news of the disaster spread so quickly around the world at that time?

Researchers are now able to understand how, by analyzing digitized historical newspapers. Understanding how the globalization of information first took place in the 19th century is the objective of the “Oceanic Exchanges” (OcEx) project, which unites researchers from the USA, Mexico, Finland, the United Kingdom, the Netherlands and Germany. At the University of Stuttgart, the ILW and the Institutes of Natural Language Processing (IMS) and for Visualization and Interactive Systems (VIS) are involved.

The mass media was made possible in the 19th century the printing press and telegraphy. And more and more people were able to read, which increased the demand for political, but also sensational news. “Newspapers were the Facebook of the 19th century”, says Priewe, quoting Ryan Cordell, the initiator of OcEx. The professor of American literature at Northeastern University in Boston has been investigating how the various newspapers copied one another for years. “For example”, Priewe explains, “a news story would start in New York and show up a few days later in Louisville, Kentucky.

“Some new stories ‘went viral’, as we would say today – especially news about notable persons.” Even in those days: “more was reported about a politician’s family circumstances and their choice of clothes than anything else”, as Priewe’s employee, the newspaper researcher Jana Keck, has discovered. In her dissertation, she investigates, how the newspapers produced for German immigrants reported on slavery among other things. “The German-language newspapers in Europe mainly reported the fact that slavery existed in the USA”, Keck explains: “The immigrants in the USA were shocked because they had not been aware of the existence of slavery. “What we see is that the German-language newspapers in the USA created a platform for discussions about the contradiction between the American emphasis on liberty on the one hand and slavery on the other.”

We want to make the data digitally available to both the public and the science community and to save the newspaper pages from decay”

Jana Keck, Department American Literature and Culture at the Institute of Literature

The data used in the OcEx project is based on over 100 million computer-readable newspaper pages from more than seven countries. According to Priewe, not even ten per cent of historical newspapers in Germany have been digitized. The process got underway in the 1990s and, to this day, libraries are still scanning their collections. For many years, their efforts were uncoordinated, which resulted in the production of inconsistent data sets. “We want to make the data digitally available to both the public and the science community and to save the newspaper pages from decay”, says Keck. This is why the Stuttgart-based OcEx team is collaborating closely with the Berlin State Library, which publishes guidelines and trains other libraries in best practices for digitization, as the quality, particularly of early digital products, is often poor.

“Text recognition software sometimes produces some weird character strings”, says Keck. That is why the group in Stuttgart is developing sophisticated digital search and word processing tools. “In this way we get robust results because we no longer search for single words, but for whole phrases.” The team uses neural networks in probability models to predict which words will occur in which context, for example, when “lock” refers to a door lock and when it means a canal lock. The researchers are also developing ways pf presenting the research results in a visual and comprehensible way.

Prior to digitization, one had to visit the archives and go through every single newspaper issue. We’re at the start of a revolution, now that we can search the data by entering keywords on the computer.”

Prof. Marc Priewe, Head of Department American Literature and Culture at the Institute of Literature

OcEx’s tools not only render the ways in which news was spread visible, but also show who changed the respective texts. As Priewe emphasizes: “One cannot overestimate the importance of these possibilities. Prior to digitization, one had to visit the archives and go through every single newspaper issue. We’re at the start of a revolution, now that we can search the data by entering keywords on the computer.” In another example, Native Americans in the southeastern USA were forcibly relocated to Oklahoma in the 1830s. “Historians refer to this episode as the Trail of Tears. Searching for this term in contemporary newspapers won’t produce any results” – because it was only coined at a later date. “We’re working on a tool that will first look up the Wikipedia article on the Trail of Tears and then the contemporary newspapers with search terms gleaned from that article.”

Although the OcEx project will end in the summer of 2020, “the humanities questions will never stop”, says Keck. Thanks to digital tools, it may be the case that nothing will be as new in the future as the newspapers of yesteryear.

Prof. Marc Priewe
Head of Department American Literature and Culture
Institute of Literary Studies

Contact

 

University Communications

Keplerstraße 7, 70174 Stuttgart

To the top of the page