“To mine and to tie – Text mining and network analysis for historians”
This was the title of the 9th Historical Network Research Workshop, which took place at Ruhr-University Bochum, organized by the SeNeReKo team. More than 40 participants to the Ruhr Area to take part in programming introductions and other workshops, to listen to speeches by their colleagues from different research fields, having a network speed dating, and discussing lively, but in relaxed atmosphere the past, present and future of network research and Digital Humanities.
Started as an informal gathering of historians in 2009, the Historical Network Research Workshop became a growing annual meeting not only for historians. It happened to be the SeNeReKo research team organizing the 9th HNR workshop in Bochum from April 10th until April 12th. Not only were the religious scholars of SeNeReKo not historians In a total nearly 50 international scientists took part in the workshop. Nevertheless, much “real-life” social networking was done. Many contacts were established between the panels, which preserved the informal, relaxed atmosphere.
Activities like a guided tour through Bochum, a common dinner or a network speed dating contributed to this atmosphere, too. More networking was done in the internet: During the conference, the hashtag #hnrws2015 was used to send more or less scientific information about it to the world.
Keynote speech: Solution to real world problems by network analysis
The official beginning of the workshop took place on Friday at 4 p. m. After an introduction by CERES (Centrum für Religionswissenschaftliche Studien, Center for Religious Studies) scientific coordinator Beate Hofmann and Martin Stark of the Historical Network Research platform, Jana Diesner held her keynote speech about “Developing Practical Solutions to Real World Problems by Going from Words to Networks”. The main topics were how to construct network data from text data and pertaining meta-data, and how to
jointly consider text data and network data for analysis.
The network researcher from the University Illinois Urbana-Champaign (UIUC) presented how network research can be used to evaluate the impact of real world problems – and foundations’ countermeasures – on the basis of a database of newspaper reports and social media activities. At this point, Diesner shortly introduced the network analysis software ConText which she and her team developed. The next day she gave an introduction to this software.
Before the official beginning two members of our SeNeReKo team, Frederik Elwert and
Sven Sellmer, offered interactive introductions to the programming languages Python & Gephi and R respectively. While R is well known for its statistical and graphical capabilities, Python is well suited for working with text, XML, and networks, making it a good candidate for historical network analysis. Both workshops revealed that one does not have to know programming languages perfectly in order to digitize humanities.
Take words for the nodes, not persons
Elwert also held another workshop on Friday morning, where he explained how to understand analyzing a text as a network. One can take “Words for the nodes of the network, not persons. Relations between words are inferred from shared context, or from similar usage patterns across the text. Networks can then be used to analyze and visualize text structures, semantic relations, or fields of meaning” he writes. This is an approach we use at the SeNeReKo project.
Andreas Kuczera (Justus Liebig University Gießen) asked a rather provocative question: “– that was the title of his presentation. The historian spoke about his first doubts with network visualization, but he made good experiences with this method during his work with the digitized Regesta Imperii, a collection of regests of the Holy Roman Empire. Kuczera
showed “how looking on graph based networks can open up new and surprising perspectives for historical research”.
He was followed by Silke Vanbeselaere’s workshop about “Digital Prosopography and Network Analysis”. The graduate student from the University of Leuven demonstrated what a practical tool Gephi is in order to analyze historical groups of persons. By the means of network analysis one can easily “identify and disambiguate historic actors in a large dataset.”
Historical Network Research 2.0
No scientific gathering without theoretical considerations: Daniel Reupke held a speech about some “Thoughts on Past and Future of a Fashionable Research Approach” (“HNF2.0?! – Überlegungen zu Vergangenheit und Zukunft eines modischen Forschungsansatzes”). Ten years after network research found its place in the historian’s toolbox, Reupke showed some future perspectives of this method. With the example of a network based on the correspondence of composer Georg Philipp Telemann he presented HNR 2.0, by which one can e. g. identify persons automatically.
“Canonical citations are the standard way of citing primary sources in Classics”, writes Matteo Romanello of the German Archaeological Institute. “Such citations play a key role as they signal, in journal articles and other types of secondary sources, text passages that were studied in relation to one another.” In his presentation “Transforming Indexes Locorum into Citation Networks” he showed how networks not only exist between persons (or words, as we have seen before) but also between texts.
Another word-based presentation came from University of Heidelberg’s Istvan Czachesz: “The Study of Word Co-occurrence Networks in the Greek New Testament” was the title of his speech, in which he took into consideration not only the linguistic features of the biblical greek dialect, but also the cultural context of the first century Mediterranean and the underlying psychological model of textual reception.
Gout: Same word, same thing through the times
There can be found 20 terms for gout, also known as arthritis, in greek and latin scripts from the 5th century BC to the 15th century AD. Dariya Rafiyenko and Hannes Kahl found 1570 tokens of these types in their corpus. These both scholars from Leipzig demonstrated in their speech “Gicht und ihre Formalisierung” (“Gout and its
Formalization”) how one can make statements regarding historical phenomena by the means of statistics.
The last talk of the workshop, “Folklore and Social Media”, was given by Mari Sarv and Liisi Laineste from the Eesti Kirjandusmuuseum (Estonian Literary Museum). The underlying thesis of their research on folklore is that “, although the media used for communication have diversified in the present days, the mechanisms that regulate interaction and the functions it carries have largely remained the same” through the ages. They showed that “intertextuality should thus be highlighted as a central feature that lies at the crossroads of the medium, the people and the message.”
Sunday’s special feature was the network speed dating. Seven experts – the presenters – opened up to small groups of other participants and their remaning questions about projects, methods and contents. Both sides, the experts as well as the ones with the questions, only had a few minutes to discuss their problems. When the time was over, the
expert was available for the next curious guest.
Honestly, we did not expect that the communication among the participants would be so good, so the speed dating had less participants than we thought – all questions seemed to have been answered at lunch or during the coffee breaks. As organizers we are not sad about it. Taking the feedback we got into consideration, it was a thriving workshop for our guests and for us.