A Perfect Union of Contrary Things: A Collegial Polemic on DH Infrastructure

Terhi Nurmikko-Fuller

RMIT University, Australia · terhi.nurmikko-fuller@rmit.edu.au

J. Stephen Downie

University of Illinois Urbana-Champaign, USA · jdownie@illinois.edu

TopicsEnglish meta-criticism (reflections on digital humanities and humanities computing)Library & information science

Keywordsinfrastructure FAIR CARE FATE

As a discipline, the Digital Humanities (DH) have long engaged in the process of building and developing infrastructure at various granularities. Earlier examples tend to be more mechanical (e.g. punchcard systems), with more recent examples consisting not of individual machines or clusters of machines but of international networks of colleagues, collaborators, and institutions. Initiatives such as Europeana
Europeana: https://www.europeana.eu/en
and DARIAH-EU
DARIAH-EU: https://www.dariah.eu/
and NovelTM (Underwood, et al., 2020) show that platforms can and do easily span across international borders, language barriers, and jurisdictions. Even at a smaller scale of infrastructure projects, examples such as Omeka
Omeka: https://www.omeka.net/
, Knowledge Commons
Knowledge Commons: https://hcommons.org/
, and Transkribus
Transkribus: https://www.transkribus.org/
show that DH research tools with even very different business models and technological implementations can be successfully applied in various different areas within (digital) Humanities research. Yet, many of these structures are evaluated and used exclusively in those project-specific contexts. We argue that infrastructure can, especially when built ethically and with an aim to common benefit, present us with opportunities for bridging the discipline on a global scale, far beyond the islands of individual projects.

Our paper derives its title from a New York Times Bestseller: A Perfect Union of Contrary Things (Jensen, 2018). It is the authorized biography of musician and vintner Maynard James Keenan, who has been involved in a number of rock bands, the most famous arguably being Tool. Keenan’s lyrics are often centred on identity, self-reflection, and understanding. It is with these lofty goals in mind that we, who have forged our careers as practitioners of the DH, tackle the questions as to the big picture take on the current and future challenges facing digital infrastructure in our discipline. What follows is more of an opinion piece than a project report, more reminiscent of a letter to the editor than a summary of a project with a workflow and results. It is a call for engagement, an invitation to inspire the DH community to build the types of digital infrastructure that can truly bridge global, disciplinary, and methodological divides.

Building infrastructure in DH is not new. Funding cycles for entities such as the Australian Research Council
Australian Research Council: https://www.arc.gov.au/
or Horizon Europe
Horizon EU: https://enspire.science/horizon-europe-funding-mechanism/
are typically geared towards projects with an intended run-time of a maximum of five years. This means that innovation and the creation of new things are rewarded and supported – the maintenance of existing tools and resources less so. This leads to an academic culture that promotes the rapid prototyping and building of new infrastructure for most disciplines, including DH and its adjacent fields. Examples of longevity come in the form of standards such as the Text Encoding Initiative (TEI)
Text Encoding Initiative: https://tei-c.org/
and projects such as the Pleiades ancient world gazetteer
Pleiades: https://pleiades.stoa.org/
, the Github pages for which were set up 12 years ago.
Pleiades Github: https://github.com/ryanfb/pleiades-static-search

Whilst the building of DH infrastructure may be a well-established methodology in the discipline, discussing said infrastructure in the context of responsible computing might just be a new angle. Although recent work indicates there is a growing appreciation for the need for forward-thinking strategies in the development of DH infrastructure (Ross, J., et al, 2024), many papers fall into one of two categories: project papers, which report on short-term, project-specific platforms, databases, or websites as “digital infrastructure”. The second, panel discussions, such as the “Text Analysis Tools and Infrastructure in 2024 and Beyond” published in the proceedings of the ADHO DH2024 conference. We want to open up the concept of “infrastructure” as something that transcends specific domains of expertise, and particular pieces of software, even named methodologies. DH infrastructure should represent the diversity of DH as a discipline and exist to support DH research and scholarship across methodological and domain-specific boundaries.

The discipline seems to lack a robust definition for “infrastructure”. Rarely, for example, have papers included human contributors (software developers, data scientists, or specialist researchers), even if they play an irreplaceable role in a given project workflow. Few papers propose approaches for proposals that explicitly make infrastructure robust to future change. This seems an ethical issue, and perhaps a current weakness in the way we as a discipline recognise and reward contributions to projects and publications.

In (Data) Principle

How should we approach this relatively new way of discussing infrastructure? As practitioners we can look to existing examples from relevant contexts. The driving force behind DH infrastructure could be derived from responsible computing, whilst proposing the need for a holistic understanding of the entire process of developing software. The idea calls for frameworks that can be applied to ensure that technology is both developed and used ethically; that solutions are proposed for ecologically and economically sound sustainability; that tools, algorithms, software, are all built with the explicit aim of societal benefit; that worse-case-scenarios are considered and potential for harm is minimised. More ambitious aims include green IT (environmentally sound approaches) and take socio-cultural and political issues into consideration (anything from fairness to privacy and inclusivity); in governance, responsible computing promotes transparency and accountability.

Those involved in the Open Data movement will be very familiar with the FAIR data principles,
FAIR data principles: https://www.go-fair.org/fair-principles/
and their call for data to be Findable, Accessible, Interoperable, and Reusable. These ideas play easily into ideas of responsible computing, and even if not explicitly articulated as such, are the starting point of many a DH project. The First Nations’ response to FAIR, in the form of the CARE data principles,
CARE data principles: https://www.gida-global.org/care
is a stronger parallel still, calling for Collective Benefit, the Ability to Control data, and to make sure it is collected, analysed, stored, and accessed Responsibly, and Ethically. These concepts are derived from the core principles of Indigenous Data Governance, which we recognise and acknowledge is not equivalent to the governance of Indigenous data (that many tertiary education and cultural heritage institutions in the Global North carry out).

The overlap between these data principles and responsible computing comes in the form of the FATE principles (Memarian & Doleck, 2023): Fairness, Accountability, Transparency, and Ethics. The challenge here is that these principles, in order to be truly effective, have to apply to all stages of a given system’s lifecycle. It requires buy-in not only from project leads, but also software developers; from those in charge of institutional policy, and national law and legislation; from philosophers, social scientists, and interdisciplinary ethicists, all of whom contribute to the wider environment and long-term sustainability of digital infrastructures.

Our first union of contrary things is thus an aggregation of acronyms: FAIR, CARE, and FATE. All are required, not all are equally delivered. The strength of the FAIR principles is that they are achievable through technological solutions, many of which are known to us already. The first three stars of Tim Berners-Lee’s Five Star Standard for Linked Open Data,
Five Star Linked Data: https://www.w3.org/2011/gld/wiki/5_Star_Linked_Data
for example, give us an international standard that ticks most if not all of the FAIR principles. CARE and FATE are a more challenging space. There are no existing tools that can program responsibility or ethics; often the driving forces for technological development are pressures and reward systems quite different from community benefit; and, ethical engagement and relinquishing authority over data can even go directly against institutional policy and modus operandi. Building relationships of trust with individuals and communities is a long-term, slow process, which can go directly against the pressures of producing technical and digital innovation at a fast pace.

Different Infrastructure Demands

The juxtaposition of the challenge between technological advancement and social benefit transcends cultural contexts. Recent DH projects such as the creation of the VR environment of Fels Cave, a UNESCO World Heritage site on the island of Lelepa in Vanuatu are a timely example of this (Nurmikko-Fuller, T, et al., 2025). They exemplify how DH projects can (and do) bridge global divides across the North and South. It also sheds light on another union of contrary things: the need for and potential of minimal computing in rural locations and for cultural heritage, with the expensive and significant long-term storage needs of raw data that cannot be delivered on location. In short, it is an example of a project where the lack of digital infrastructure on Lelepa can be contrasted with significant digital infrastructure facilities in the academic institutions who collaborate with the Indigenous community on this project. In this case, digital infrastructure looks very different on Lelepa, consisting of mostly hand-held devices, laptops, and cameras. These enable the capture of the site using minimal computing methods such as photogrammetry, but the perpetual storage of the 2D digital photographs that constitute the “raw” data that is used to create the 3D rendition of the site must fall to institutions that have extensive systems in place to facilitate the storage, and if necessary, retrieval of these photos at any point in the future. In many ways, this project epitomises the type of “infrastructuring” in DH as described by Pawlicka-Deger (2022); it was a process of co-creation, participation, community engagement, and a shared interest in the preservation and research of Fels Cave – and it would not have been possible without the long-term relationships established and maintained by anthropologist Christopher Ballard with the community on Lelepa Island.

Looking forward, the next pairing of contrary things is the challenge presented by the need for scaling, and for scaling securely, when faced with exponential growth in demand. Furthermore, the exponential growth of digital infrastructure in the form of data centres has an immense, potentially detrimental impact both ecologically (Barringer, 2025), and economically, affecting societies or parts of societies in potentially uneven ways. As practitioners in DH, as builders of new and ever more complex digital infrastructure, and, as educators of up and coming generations of software developers, policy makers, and users of digital infrastructures, we need to begin to address these challenging aspects. Digital infrastructure has increasingly great demands on electricity and water – these impact not only those contexts where members of society foot the bills with ever increasing utilities, but also those far beyond the reach of the benefit of much of what these data centres have to offer.

Concluding Remarks

Our opinion piece is not a rage against the machine, but rather an attempt to flag that all this development of digital infrastructure does not take place in a social or environmental vacuum. We pose the questions whether the processes of creating ever more complex and data-hungry systems to perpetuity is manageable and maintainable. As a discipline, we need to ensure that what we build now benefits societies more widely beyond specific research clusters or national systems. We need to ensure there is benefit not just in the immediate, but the long-term future as well.

This paper represents just the first tentative steps into a complex space. We look forward to engaging our colleagues at DH2026 in discussing all things infrastructure. It is through engagement that we can build the type of sustainable, beneficial infrastructures that can help us bridge across global divides.

References

Barringer, F. (2025). “Thirsty for power and water, AI-crunching data centers sprout across the West”. & the West. Bill Lande Center for the American West, Stanford University, Palo Alto, California.
Jensen, S. and Keenan, M. J. (2018). A perfect union of contrary things. Rowman & Littlefield.
Memarian, B., & Doleck, T. (2023). Fairness, Accountability, Transparency, and Ethics (FATE) in Artificial Intelligence (AI) and higher education: A systematic review. Computers and Education: Artificial Intelligence, 5, 100152.
Nurmikko-Fuller, T., Nelson, K., Ballard, C., Wilson, M., Farenearu, R. L. M., and Wille, E. (2025), Digitising Fels Cave, Lelepa Island, Vanuatu. In Proceedings of the international Digital Humanities conference 2025 (DH2025), Lisbon, Portugal, 14 – 18 July.
Pawlicka-Deger, U. (2022). Infrastructuring digital humanities: On relational infrastructure and global reconfiguration of the field. Digital Scholarship in the Humanities, 37(2), 534-550.
Ross, J., Sheail, P., Terras, M., & Schofield, C. (2024). Infrastructure Futures for Digital Cultural Heritage. Zenodo. https://doi.org/10.5281/zenodo.13710266
Underwood, T., Kimutis, P., & Witte, J. (2020). Noveltm datasets for english-language fiction, 1700-2009. Journal of Cultural Analytics, 5(2). https://doi.org/10.22148/001c.13147 .

PreviousQuest for Unrelenting Experimentation of Durable Electronic Editions [Iteration 2, Implementation 1]NextBetween Collections and Data: Engaging GLAM Institutions in a Federal Open Science Ecosystem through FedOSC