What is the role of the Arts and Humanities in the age of Data Science? Two short proposals
In September 2018, I participated in a panel that explored the question ‘What is the role of the Arts and Humanities in the age of Data Science? The digital dimensions of this question are addressed on an ongoing basis by the Turing Institute’s Data Science and Digital Humanities Special Interest Group, of which I am a member. The Edinburgh panel was supported by that group and Digital Scholarship at the University of Edinburgh. I had 5 minutes to speak and so I presented two short proposals. Here is an approximation of what I said (or aimed to say):
- The Humanities can play an important role in the building of a ‘Critical Data Science’
What do I mean by the term ‘critical data science’? Essentially, I am referring to a data science that foregrounds questions of power, culture, context and ethics and seeks to understand how they shape, and are shaped by, the data-driven methodologies that data scientists develop and use. As a wealth of recent scholarship is showing, neither data, nor the computational techniques that are used to wrangle data, are necessarily neutral. Safia Umoja Noble’s Algorithms of Oppression, for example, has shown how search engine results can misrepresent and discriminate against women of colour. In turn, these results can reinforce racism and sexism in wider society. The disciplines of the humanities, with their focus on how context, culture and power shape the construction and definition of knowledge, and knowledge-makers, can thus play a vital role in the building of a critical data science. My colleague Tessa Hauswedell and I have been discussing the kinds of questions that can be asked to evoke the many ways that the humanities can and is contributing to the multidisciplinary project of building of a critical data science. They include questions like ‘how can we devise new approaches to search, mine, extract and evaluate the textual, visual and audio data contained in digital archives that are globally accessible, but always bound to specific cultural and institutional contexts?’ And ‘how can the analytical and interpretative approaches of the humanities foster deeper understandings of how socio-cultural contexts shape epistemologies of data science?’
- The Humanities can speak truth to power but must not shirk the necessity of holding itself to the same standards
An essential preliminary stage of doing many kinds of data-driven analyses involves the cleaning of data. Sometimes it is possible to automate this work, using approaches from fields like natural language processing. When it can’t be (fully) automated, a human will need to do what is often repetitive and laborious work. Data science discussions and hand books often point to market places where individuals can be engaged to do this work at shocking low rates of pay. Recently, an article in the Atlantic has argued that the internet is giving rise to ‘a new kind of poorly paid Hell’. This is all the more troubling when it is contextualised with regard to the longer history of computing, as Mar Hicks has argued in Programmed Inequality: ‘… computerization is an explicitly hegemonic project built on labor categories designed to perpetuate particular forms of class status.’ (p.6).
So, my question is about the role and responsibility of the disciplines of the humanities in pushing back against such inequities and disrupting such power dynamics. Can the humanities help to foster a critical data science that does not devalue contributions that may be quotidian but are essential to actually getting work done? How can a critical data science ensure that those who contribute to this work are fairly remunerated and treated with dignity?
When thinking about this in connection with data science, or the application of data science to humanities research, we cannot shirk from shining the spotlight on the humanities itself. This is uncomfortable to do because whether in the historical or present-day context it is not difficult to find examples of how the contributions of some categories of workers to the humanities and digital humanities have been devalued. In the digital humanities, for example, this has sometimes resulted in the overlooking or silencing of the contributions of those who do technical work and research. In ‘No Job for Techies: Technical Contributions to Research in the Digital Humanities‘, Bradley, for example, has alerted us to the way that the diminutive term ‘techie’ can serve to normalize divisions between those on academic and technical tracks. Projects like the Collaborators’ Bill of Rights are striving to address issues related to this but it is hardly a revelation to say that the (digital) humanities, and the disciplines and professions with which it collaborates, must do better in this regard. In this way, the humanities has much to give to, and take from, the age of data science. Thank you for your attention.