Uncertainty plays a role on various levels within Digital Humanities research. The fourth Digital Academy aimed at discussing methods and theories to deal with uncertainty as well as benefits and challenges that result from it. Further, on the Open Space Day, three experts gave insights into their strategies on how to face uncertainty within their academic research.

The main focus lay on the incompleteness, ambiguity and uncertainty of the data which we work with in the Digital Humanities. MARCELLA TAMBUSCIO (Graz) made a start by outlining uncertainties faced within historical network science and prosopography, exploring various sources related to Maximilian I (1493–1519). She pointed to the challenges interpreting incomplete and biased data as well as the collection and modelling of such data. Working with network analyses, she points out that this is often used for visualisation purposes only and shows how this in fact can be efficiently applied to the conduction of prosopography studies, too. MARIJA BLAŠKOVIĆ (Barcelona) gave an insight into creating a database based on historical data on women as underrepresented individuals in medieval sources in the field of Iberic Studies. Here, researchers often tend to investigate minority groups through close reading methods, whereas when exploring such data through digital methods, we do face additional challenges of disambiguation, for example the standardization of names or identification of individuals.

SHANMUGAPRIYA T (Toronto) focuses on uncertainty in the context of working with mixed methods and tools to explore historical data from the global South during the 19th and 20th centuries. Aiming at a historical gazetteer by extracting toponyms from a text corpus sourced from the British Colonial India Corpus, she demonstrated issues when trying to identify toponyms with computational methods such as Named Entity Recognition, since the sources often are intricate and provide non-standard formats and incorrect spellings of toponyms. FLORIAN THIERY (Mainz), too, gave an insight into facing uncertainties in the field of georeferencing. Besides pointing to the complexity of creating data sets from uncertain and ambiguous data – e.g. due to vague location descriptions or different granularities – he also addressed the challenge of modelling uncertain data in a way that is commonly understandable across different disciplines.

It was thus found that uncertain data plays a vital role in the Digital Humanities and poses several challenges. However, not only does the usage of digital methods within humanities research evoke difficulties and limitations as described above, but it can also help resolve problems and, most importantly, disclose uncertainties. One of the main questions remains how to represent and visualize uncertain data. Dealing with data normalization, enrichment and georeferencing, CATRINA LANGENEGGER (Basel) reported on difficulties when processing and organising uncertain data, too, working with fragmented reports of Swiss Refugee Camps during the 1940s. Here, she particularly faces incomplete data as well as misspellings. In order to represent degrees of uncertainty when it comes to incomplete data, she designed a system of codification. However, the visualisation of uncertain data still poses issues within academic research, not least because of software restrictions and technical limitations. Complementary, VALENTINA PASQUAL (Bologna) presented her work on representing uncertain and subjective claims on provenance and authenticity within knowledge graphs in order to display the multifaceted domain of critical discourse in the humanities. She argued that the uncertainties she faces in her research can be addressed by adding weaker logical status claims (WLS) to the data sets, working towards a more comprehensive and dynamic approach to complex information within knowledge graphs.

ALBERTO CEVOLINI (Bologna) and ELENA ESPOSITO (Bielefeld) presented a different approach to uncertainty. They investigate how shared uncertainty in the context of the insurance industry can be conceptualized with regard to the disruption through digital technologies and individualization processes. They found that shared uncertainty is exploited by the insurance industry. New digital technologies enable the collection of individual behavioural data, which affects the role of uncertainty within insurance systems as well as the role which insurance plays in society. Additionally, they added the question of terminology and emphasized the importance of conceptualising and differentiating terms related to uncertainty within varying contexts, for instance risks and dangers. Eventually, FRIEDERIKE WINDEL (New York) offered an insight into the various levels of uncertainty which she encountered during her research on analyzing emotions within social media data. She particularly focuses on the concept of Heimat with regards to the construction of whiteness. Beyond facing challenges with uncertain data, theories and methods, she also drew attention to the question of how to present uncertainty within academic texts as well as the dangers of fast-moving and fast-changing web platforms that we might source our data from.

After these contributions, Silke Schwandt (Bielefeld) interviewed three experts in the field of Digital Humanities on their encounters with and perspectives on uncertainty. MICHAEL PIOTROWSKI (Lausanne) first highlighted the importance of being aware of uncertainty during the own research process and making it explicit. His concept of historiographical uncertainty involves that historians construct narratives which are based on historical data and information but also go beyond - causal relationships are created between people, places and events. He understands doing history as modelling: It is only because we do not know everything about the past that we can write history. History, therefore, is not the same as the past, but it is the way we make sense of the past. This is because there is an innate desire for explanations and answers – and therefore reducing uncertainty. Working with machines, technical tools and methods can certainly lead to new uncertainties. However, Schwandt and Piotrowski agree that in this context, dealing with digital data and digital methods also forces and helps us to be more explicit in our research when modelling. We need to be precise when it comes to definitions, conventions and standards.

JANE WINTERS (London) put emphasis on the fact that with different kinds of data come different kinds of uncertainties. Focusing on web archive studies, she encounters challenges such as changes in the accessibility of sources and lacking transparency of data collection processes. She admits that the web archive studies are still in the early stages and that searching through these archives is still like rummaging in a large data bucket. Moving away from natural language search, she calls for a more structured approach, utilising metadata and the utilisation of metadata, links and hierarchies. Moreover, she states that there is an inherent linkage between the Digital Humanities and uncertainty. The data usually reflects human experience and therefore is complex and uncertain. She highlights that our task is not to look for truth and facts but for possibilities and interpretations. The preservation of data, too, adds a layer of uncertainty: Digital data is as fragile and vulnerable as data stored in physical archives, which makes it imperative to be aware of the ways archives are hosted and data collections are shaped. However, not only is the data we deal with vulnerable, but so are the digital tools and equipment we use. This poses obstacles and challenges that are of interest to various disciplines, not only within the humanities but also the information and computer sciences. At the same time, there are manifold opportunities arising from uncertainty: Most importantly, new research questions are generated from being uncertain. Thus, it is crucial to be sensitive to as well as allow for uncertainties rather than glossing over them, for we will then be able to create new narratives. After all, Silke Schwandt and Jane Winters agree that teaching tolerance towards uncertainty and being transparent about the ways uncertainty interacts with our research methods increasingly gains importance.

Regarding uncertainty, SUSAN SCHREIBMAN (Maastricht) particularly focused on the presentation of research processes and results. In a digital setting, we are able to present outcomes in new ways other than just text, which also means that we can use digital formats in order to point to uncertainties and make them more transparent, for example by creating multiple versions of 3d-models as different reconstruction interpretations. At the same time, transparency remains an issue when it comes to publishing research outcomes. Not only is there an increasing need for documentation of research processes, but also for the disclosure of the decisions made when processing sources and creating data sets. Therefore, she encourages researchers, who publish their findings, to focus on how to make their decisions and research processes transparent and accessible, too. Further, we have to be aware that the data sets themselves are always incomplete and uncertain, therefore we can only make claims on the basis of a limited amount of data. Eventually, Silke Schwandt and Susan Schreibman accentuated the need for teaching data skills in order to reduce uncertainty when dealing with digital methods and data as we constantly face new technological changes, and we need to conceptualise them and find ways to effectively use them in the field of the Digital Humanities.

Overall, the Digital Academy 2023 demonstrated the importance of acknowledging, allowing for and consciously engaging with uncertainty during the process of data-driven research as well as the challenges and pitfalls of dealing with uncertainty. If we manage to locate uncertainty and find where the gaps are, we can gain new perspectives and research questions from that. In this context, the Digital Humanities can both provide methods and solutions to face and communicate uncertainty and uncertain data and also create new uncertainties at the same time. Constant changes in technology do not only cause constant new challenges but also solutions. To face uncertainties, it is crucial to embrace and improve interdisciplinary work. This way, uncertainty on various levels can be perceived as beneficial to our research rather than a threat.

Conference Overview:

Introduction: Silke Schwandt (Bielefeld) and Christian Wachter (Bielefeld)

Panel 1
Moderation: Anna Neubert (Bielefeld)

Marcella Tambuscio (Graz): Network Science to Face Uncertainty and Incompleteness in Digital History and Prosopography

Catrina Langenegger (Basel): Where was it...? Normalization, Data Enrichment and Georeferencing in the Research of the Refugee Camps of the Swiss Territorial Service 1942–1946

Alberto Cevolini (Bologna) / Elena Esposito (Bielefeld): Coping with Uncertainty in the Insurance Industry: The Disruptive Power of Behavioural Data

Panel 2
Moderation: Sophie Spliethoff (Bielefeld)

Shanmugapriya T (Toronto): Recalibrating Data Methods and Tools for Non-standard Historical Data from Global South

Valentina Pasqual (Bologna): Representing Uncertainty and Competing Claims in Cultural Heritage Knowledge Graphs: A Conjectures-based Approach

Marija Blašković (Barcelona): Embracing Incompleteness: Medieval Historiography from the Gender Perspective

Panel 3
Moderation: Christopher Kuhlmann (Bielefeld)

Florian Thiery (Mainz): Dealing with Doubts in Sites Georeferencing in Archaeology and the Geosciences

Friederike Windel (New York): Uncertainty in Analyzing Emotions on Twitter in a Mixed Methods Study of Heimat, Whiteness and Nationalism

Open Space Day – Interviews and Subsequent Discussions
Interviewer: Silke Schwandt (Bielefeld)

Michael Piotrowski (Lausanne)
Susan Schreibman (Maastricht)
Jane Winters (London)