One of the hallmarks of humanities research infrastructure is broad accessibility and applicability. If we think of libraries as the quintessential humanities research platform, even the most specialist libraries are made not for a specific purpose, but for the widest possible range of inquiry — and for public as well as expert uses. One of the chief attractions of the digital humanities has been the opportunity for work that is public-facing, whether offering new formats for analytical and interpretive scholarship, or building new resources. Those new open resources include all manner of digital materials, including data.
In 2022, I spoke with Daryle Williams about the Journal of Slavery and Data Preservation, an initiative to publish (and thus preserve) peer-reviewed data sets. Today, I speak with Barbara McGillivray about the Journal of Open Humanities Data, which “features peer-reviewed publications describing humanities data or techniques with high potential for reuse.”
Barbara, could you speak to the complexity of what we still sometimes refer to collectively as “digital humanities,” and some of the many diverse forms that it now takes?
The label “Digital Humanities” (DH) is extremely broad. It includes text encoding and digital editing, computational text analysis and corpus linguistics, network analysis of historical social relations, digital art history, digitization and metadata work, critical and cultural analytics, and infrastructure-building. Some of that work applies computational methods to long-standing humanities questions; some studies digital culture itself; some is closer to data engineering and infrastructure matters in cultural heritage.
The diversity is institutional and generational too. Some scholars sit in dedicated DH centers or departments (as in my case); others do computational work inside traditional departments like computational linguistics inside linguistics or digital history in history. A growing number of people in libraries, archives, and museums do digital humanities as a core part of their job, not a side interest.
Do you think that non-specialists (either in humanities generally or even people who are familiar with computational social scientists) are still surprised by the extent and intensity of data-driven work in the humanities?
I think so, definitely. Humanities scholars often underestimate how much of the work we do rests on shareable data, such as annotated corpora, digitized manuscript, reconstructed social networks from the past, or documentation for endangered languages. On the other hand, in my experience computational social scientists are often surprised by the nature of humanities data. They may associate data primarily with surveys or web-scraped content, rather than datasets tracking, for example, historical records or ancient fragments. None of this is actually entirely new, but large language models have dramatically increased both the scale and pace of such work, radically changing many people’s perceptions of what humanities research now involves.
How does The Journal of Open Humanities Data enter into this complex field of digital humanities, and why is “open” an operative term for the scholarship you publish?
The Journal of Open Humanities Data (JOHD) occupies a distinctive place in the publishing landscape because, for us, the scholarly contribution is the data and the methods used to create it. In much humanities scholarship, datasets remain largely invisible. At JOHD, we turn that around: the dataset and the decisions involved in creating, curating, and documenting it are the primary objects of peer review and publication.
I’ll admit that the challenge looked very different in 2019, when I took over a journal that had published just one article that year. Since then, we’ve treated JOHD as more than a venue for submissions. We’ve developed a largely early-career volunteer team that manages our presence across X, Instagram, YouTube, WeChat, and LinkedIn. We’ve also created spaces for the community to come together through events such as the Open Humanities Data Forums and, most recently, our 10th-anniversary conference, which brought together more than 50 attendees and 40 speakers from institutions across four continents. We’ve also reflected on our own role in the field, publishing research on the impact of humanities data papers and the value they create. All of these activities are part of the same broader effort to demonstrate that open humanities data is not just useful, but essential, and that the work of building and sustaining it deserves recognition and support.
Could you describe how the requirements for publication (structure of data sets, for example) in the JOHD facilitate scholarly exchange in specific ways? And related, why there are two different types of pieces you publish?
Data papers focus on making research data reusable and citable. These papers require authors to deposit the underlying data in an open repository with a persistent identifier. For example, Mohsine Aabid, Simon Dumas Primbault, and Patrice Bellot published a data paper about a dataset of user navigation sessions and content metadata from a digital library, reconstructed from server logs and has potential for reuse in areas such as scholarly browsing behavior, recommendation systems, and digital library design.
We also publish discussion papers, which provide space to explore issues that go beyond the description of a dataset itself, such as methodological challenges, questions around data sharing and reuse, and broader debates relevant to research data in the humanities. For example, Erdem Uçar’s paper addresses a long-standing challenge in the study of Old Turkic texts: over the past century, scholars in different countries have used different transcription conventions, making editions difficult to compare. The paper proposes a framework that preserves each edition exactly as it appears while also creating a standardized version, enabling researchers to search and analyze texts consistently across different scholarly traditions.
Could you speak to a couple of the most recent pieces, one of each of those types and share how those scholars and their research reflects and advances humanities scholarship?
We’ve learned a great deal along the way, much of it through experimentation. This has involved developing new editorial processes to manage the growing scale of manuscripts we receive. We are also exploring ways of integrating more explicit forms of data peer review into the publication process and carry out research on issues relevant to scholarly publishing, such as the automatic linking of data papers and relevant research papers using Natural Language Processing techniques or the management of the journal itself. At the same time, JOHD has become a space for training and mentoring early-career researchers, many of whom have taken on increasing editorial responsibilities and found opportunities to develop their careers within the journal. Looking back, the journal’s legacy is likely to lie not only in the papers it has published, but also in the community of scholars that have grown around it.
