Graph Technologies in the Humanities 2020

The conference took place at the University of Vienna on 21 – 22 Feburary 2020 at Aula am Campus, Spitalgasse 2, Hof 1.11.

The Programm can be found here.

Modelling the Scholarly Process

The idea of connecting our knowledge, or networking information, has been a desideratum of digital scholarship since its beginnings. As these networks lend themselves very easily to expression as graphs, our tools of choice have followed: many libraries, archives, museums, and other repositories of data in the humanities make their goods available via Linked Open Data, following the principles of the Semantic Web. Many other research projects in the humanities have adopted graph models and even graph databases as their base, as these lend themselves naturally to collections of information where the link, the connection, or the relationship is paramount.

The 2019 conference celebrated the breadth of application of graph-based approaches in the humanities. For 2020 we would like to follow up with a more focused theme, whose central question is: To what extent are our scholarly processes reflected in our data models, and how (if at all) do graph technologies allow us to capture these processes in the digital realm? That is to say, when we choose to link items of information together (say, the identification of a person in an inscription or of a place in a text), are we adequately capturing the process that led us to make this choice? What mechanisms are (or should be) available to enable fact checking and to reconstruct the provenance of information?

Key topics arising from the main theme Modelling the Scholarly Process  include:

  1. Provenance and source criticism – While Linked Open Data, in the form of RDF, has (outside of textual criticism) become something of a standard for repositories of humanities data, there is a certain hesitation around its reuse outside of the project in which it was created. The perception is that the RDF data will not be complete enough, accurate enough, or nuanced enough for use outside its original context. How can we address these questions of source critique, making the data maximally useful, either in an RDF-based project or in any other form of linked data?
  2. Validation – One of the points of resistance to the adoption of graph models more widely in the humanities is the lack of a validation framework as well-known as XML schemas. Insofar as validation is a necessary part of the scholarly process, how do we do this with graph models?
  3. Visualization – One of the acknowledged strengths of networked models is that they lend themselves very easily to visualization, which is a pivotal part of explaining the results and the scholarly process to our peers; at the same time, the „hairball“ of a densely-connected network graph has become something of a cliché within the digital humanities. The urgent question thus arises: which visualization methods there are beyond the “hairball”? How can visualization techniques support the analytic richness of graph complexity while at the same time reduce it to the point of comprehensibility? Which modes are there to interact with graph data and how can visualization contribute to manipulation and editing of graph data?