In the final week of October the annual IEEE eScience conference will take place in Amsterdam, the Netherlands. This will be my first eScience conference, and I look forward to seeing how the eScience community is similar to or different from the digital humanities community. I have submitted a paper applying findings of my PhD research to the development of eScience infrastructures for the humanities. Specifically, my paper discusses the problems of power asymmetry in collaborations, with scholars dependent on infrastructures developers, and of knowledge asymmetry, with scholars lacking the knowledge necessary to influence the practices of infrastructure developers. A first difference between the eScience and DH communities I already observed was in the reviews of my submission, which found my topic of interest but lacking in a (technological) solution to the problem. Unfortunately, I do not have a solution readily available, but I have extended my power relation circle with a possible way out in the development of know-how. Below you can find the abstract for my paper, and the poster I will present at the conference (designed by my wife).

Power Asymmetries of eHumanities Infrastructures

Abstract—Digital research infrastructures simultaneously en- able and confine the research practices of scholars, constituting a power relation. This power relation can be characterised as a power asymmetry, with scholars dependent on the developers of infrastructures. In order to reduce this power asymmetry, infrastructures are developed in collaboration between scholars and computational researchers. Through an analysis of 28 interviews of eScience projects in digital history I have investigated whether collaborations succeed in reducing power asymmetry. In this paper I will explore knowledge asymmetry, the ignorance of how a collaborator performs their tasks, and how this reinforces power asymmetry. I will moreover consider how these asymmetries pose a challenge in the development and adoption of research infrastructures in the humanities.

Digital research infrastructures simultaneously enable and confine the research practices of scholars. Through the availability and communication of functionalities, infrastructures shape the practices of scholars.^[1]M. Stanfill, “The interface as discourse: The production of norms through web design,” New Media & Society, pp. 1–16, Feb. 2014. The interaction between scholars and infrastructures thus constitutes a power relation, insofar infrastructures define a possible field of action for scholars.^[2]M. Foucault, “The Subject and Power,” Critical Inquiry, vol. 8, no. 4, pp. 777–795, Jul. 1982. Furthermore, this power relation can be characterised as a power asymmetry with scholars dependent on the developers to provide suitable functionalities.^[3]R. S. Koeser, “Trusting others to ‘do the math’,” Interdisciplinary Science Reviews, vol. 40, no. 4, pp. 376–392, 2016.^[4]M. L. Markus and N. Bjorn-Andersen, “Power Over Users: Its Exercise By System Professionals,” Communications of the ACM, vol. 30, no. 6, pp. 498–505, 1987. This power asymmetry is a risk for the adoption of infrastructure when the infrastructure ultimately fails to meet the needs of scholars.

One approach to reduce this power asymmetry is to develop infrastructures in collaboration between computational researchers and humanities scholars. The aim of such collaborations is to empower scholars to indicate what possible functionalities they require or desire, and coordinate with computational researchers so that both parties will be served by the process and result of development.

In this paper, I will examine whether collaborative development succeeds in reducing power asymmetry, by studying the role of knowledge asymmetry in collaborations.^[5]A. Sharma, “Professional as Agent: Knowledge Asymmetry in Agency Exchange,” Academy of Management Review, vol. 22, no. 3, pp. 758–798, Jul. 1997. This refers to the ignorance of how a collaborator performs their tasks. When a scholar is unaware of how a computational researcher contributes to a research infrastructure, it becomes increasingly difficult for the scholar to influence the work and outcomes of the computational researcher. The scholar thereby lacks the power to shape the actions of the computational researcher, putting the latter in a more powerful position.

As part of my PhD research on the coordination and translation work in digital history projects, I have conducted 28 interviews with participants of eScience projects in digital history in Belgium and the Netherlands. Participants had backgrounds in history, computational linguistics, and software development. By combining these different perspectives, I will show how knowledge asymmetry and power asymmetry played a role in these projects. Moreover, I will show that projects that were unable to reduce the asymmetries led to in- creased discrepancies between the expectations of participants and project outcomes.
For example, in one project with participants from history and computer science, the historians noted that the research environment was so unstable and slow it became unusable. One historian inquired about this issue:

What you get is that those [scholars] say ‘I want to be served’, and [the computer scientists] say ‘no, that server is for multiple experiments, you are one of the experiments, [end of discussion]’.

The historian PI of the project reflected on this as follows:

In hindsight I think [the computer scientists] should have said more about matters such as the really practical things, such as computation capacity, server space, the stability of software, how that is managed. You need money for that too. We did not have budget for that in the project, as idiotic as that seems now.

In this project, the historians were unaware of how the computer scientists would provide the technological infrastructure. They were unaware that to the computer scientists their re- search environment was just one of many experiments running on the same server. Consequently, the historians were unable to coordinate sustained performance of the environment. The computer scientists thus shaped the possible field of action of historians, in the sense that actions were confined to the limitations of computational performance. As a result of their ignorance of this shaping, the historians were unable to act upon these limitations to ensure the infrastructure would meet their expectations.

This is not to say that a power asymmetry where computational researchers hold power over humanities scholars is the only power relation imaginable. In previous research, I found that most digital humanities projects are led by humanities scholars, and consist of a majority of humanities scholars compared to computational collaborators.^[6]M. Kemman, “Boundary practices in digital humanities,” in DHBenelux, Amsterdam, the Netherlands, 2018. This suggests that humanities scholars may likewise exert power over computational researchers in a struggle of technology-push and demand-pull strategies.^[7]J. van den Ende and W. Dolfsma, “Technology-push, demand-pull and the shaping of technological paradigms – Patterns in the development of computing technology,” Journal of Evolutionary Economics, vol. 15, no. 1, pp. 83–99, Mar. 2005. However, even in this situation scholars depend on developers to implement the desired technological change.^[8]M. L. Markus and N. Bjorn-Andersen, “Power Over Users: Its Exercise By System Professionals,” Communications of the ACM, vol. 30, no. 6, pp. 498–505, 1987. For example, in another project with participants from history and software development, the historians communicated a number of user requirements to the developers. The developers technically met the requirements, yet the historians were not satisfied since the system did not work as they had imagined. One historian said about this discrepancy:

We thought [the software developers] had the experience and knew what we meant. We did not consider that it could be interpreted otherwise.

In this situation, it could be argued that the developers were unaware of how the historians desired to use the research infrastructure. As such, knowledge asymmetry was present in the opposite direction. However, the software developers remained in a more powerful position; since they had technically satisfied the requirements, it was contractually difficult for the historians to demand improvements. The developers here too shaped the possible field of action of the historians, in the sense that desired actions were confined to the developers’ interpretations and subsequent implementation.

In conclusion, knowledge asymmetry poses a challenge to the development and adoption of research infrastructures in the humanities. Both examples show that knowledge asymmetry reinforces the existing power asymmetry. With this paper, I aim to develop a better understanding of this challenge.