Out of Data

Uncertain Archives

Sam Hind: Could you say a little on what ‘uncertain archives’ are, or what ‘archival uncertainties’ entail or comprise? Are there such uncertainties that are unique to big data archives?

Kristin Veel: There are two assumptions in the terminology of the Uncertain Archives, firstly that what is known as “big data” can – with advantage – be approached as an archival problematic, and secondly that these archives are marked by fundamental uncertainties. Our starting point in the research project has been that information networks raise questions that are familiar to the theoretical regime of the archive developed by poststructuralist thought – including philosophers such as Derrida, Foucault, and Deleuze and Guattari – who in different ways challenge the authority of the archive as a reliable repository, their capacity to produce truth, offer evidence, police experience, categorize the identities of subjects and exhibit an aura of authenticity. These theoretical approaches are helpful also for thinking about big data, we contend.

Although the marketing rhetoric around big data often focuses on how more data can bring greater and more objective certainty, we in this research group are interested – as others are in the emerging field of Critical Data Studies – in the uncertainties that also arise with the aggregation of large amounts of data, such as new biases, new forms of systemic errors and new ethical dilemmas. We focus on unknowns/unknowables, errors, and vulnerabilities as three “types” of uncertainty that are borne out of thinking about big data as an archive. What are the limits and demarcations of what is included or excluded in the archive? Which errors are tolerated and which are most feared in big data archives? Why are some bodies more vulnerable to, and potentially wounded by, big data archives than others?

SH: A lot of your work has focused on in/visibility and surveillance in contemporary culture. To what degree, then, can one say (in)visibility is negotiable? Could we, in some sense, define the contemporary digital age as that in which (in)visibility is entirely (or at least, partially) ‘non-negotiable’? 

KV: Actually, the project Uncertain Archives emerged out of conversations between myself, Nanna Bonde Thylstrup (University of Copenhagen) and Annie Ring (UCL) as part of the research network Negotiating (In)visibilities that I have been co-organising with architectural scholar Henriette Steiner. In 2015 we published the book Invisibility Studies (Peter Lang) which brought together a range of interdisciplinary perspectives on the cultural significance of how visibility and invisibility take form in urban architecture and in the surveillance technologies that pervade our everyday lives. Transparent glass facades have dominated much contemporary mid– to high–rise urban architecture since the 1990s, and we have been interested in what makes people want to live behind glass facades in what we can call a ‘contemporary surveillance culture’. This includes of course the datafication of our lives, which configures the visible in new and interesting ways, raising questions about the cultural, political and aesthetic implications of being visible. Glass is often used as a political or ideological statement that aims to signify transparency rather than necessarily embody it. Even though glass is a transparent material, it is actually only so under very specific light conditions. Transparency is thus an ideological and negotiable category. Glass thus solicits a relationship between the viewer inside and the surroundings (or the viewer outside and the interior) that gestures toward a visibility that perhaps only really exists in theory. We may think of invisibility or transparency in big data archives in similar ways. So I would not say that (in)visibility is non-negotiable, but perhaps rather that the relationship between what is visible and what is invisible should be regarded as a fluctuation or an oscillation that points to a fundamental uncertainty (to return again to that term) as to what is made visible and what is made invisible in data environments.

Timo Arnall, Internet Machine (2014)

SH: On its deliberate ambiguity, what does constructing something ‘out of data’, or being ‘out of data’ mean to you, especially in regard to your own work? How do these efforts suffer (or indeed, prosper) with the acknowledged uncertainties of (big) data archives? 

KV: I think that the framework for the lecture series that you organize here is extremely pertinent. To me ‘being out of data’ speaks to the double meaning of out as something that signifies the scarcity of a resource, but also a production of something in need of further research. I very much share the interest of the lecture series in charting the limits of data that the word out signifies – what kind of boundary or threshold are we dealing with in relation to data, what are the consequences of reaching those limits, and how does this manifest itself in, for instance, cultural forms?

Unknown, errors and vulnerabilities as the three types of uncertainties that we focus on can all be thought of within this span between the resource and the threshold that I believe that the notion of out connotes. Out is very much a spatial metaphor and in this sense a term that speaks to the senses and an aesthetic mode of reception, which I think is extremely important to address when thinking about these kinds of data environments.

In the description of this lecture series you refer to the comparison of data and oil – with all the material connotations that this entails – of seeping, leaking and of stickiness – for us in the Uncertain Archives project the inroad to thinking about the abundance of data, or data as a resource, has been the archive (Daniela Agostinho and Nanna Bonde Thylstrup are, for instance, working on an article on archival leaks), and I think the fluidity here is important and speaks to what we try to capture with the notion of uncertainty. The archives we work with here are not dry and dusty, they are liquid and leaky. The metaphorics of the oil as a liquid speaks to the risk of leaks that may render us vulnerable in manners we had not anticipated, yet data is also a resource that can be scarce and mark limits that can be used to set boundaries, build walls; thus pointing to the vulnerabilities attached to being out of data. Not being in the archive, not being registered and thereby denied access. Vulnerability is thus something that can be described as emerging out of data, but also refers to the condition of being out of data. A movement as much as condition – which is very much what we aim to capture with the term of the uncertain archive.

SH: What opportunities do you think interdisciplinary work brings to the subject of ‘constructing’ or ‘being’ (out of) data? For example, in how design, practice, theory and method/ology are, or can be, combined and considered variously? Do the systematic uncertainties and ‘negotiations’ you identify demand interdisciplinary engagement? 

KV: I absolutely think that interdisciplinary engagement is of the essence when trying to understand these kinds of issues. The question of the social implications of data-intensive environments has generated a plethora of scholarly production in a diverse range of academic fields such as media studies, art history, information science, law, geography and political science, which forms the contours of the emerging fields of critical data studies. I think such a field has to be interdisciplinary and that we need to develop new methodologies and new modes of knowledge production in order to be able to question the rhetoric of the big data archives as neutral and empirical means to organize, securitize and capitalize on data.

While media, geography, surveillance and internet governance studies have been fundamental in developing a critical framework for understanding the problematic aspects of data archives, I think it is necessary to address the sensory as much as the political, when trying to understand the archival imaginaries that we live with today. We therefore aim to take this wider field into account in the various collaborations that we engage in with computer scientists as much as contemporary art practitioners, and by bringing in distinguished scholars from across the social sciences and the humanities to inform our discussions.

Being receptive to the uncertainties of big data archives requires an attunement which asks not only what is included in the archive, who controls it and for what purpose, but also how it configures and functions through the signs, flows and bodies that populate and circulate within them, and the ways in which we meet these through narrative, sonic and visual interpretative processes. This approach constitutes a critical methodology that is responsive to archival movements and affects and offers modes of addressing these, in this way complementing the disciplines of mathematics and political science, which currently dominate much big data theorization. By engaging in collaborations with both computer technologists, artists, as well as commercial and public institutions that deal with these issues on an everyday basis, I think it is possible to bring forward a theoretically-informed approach; both to the technical and to the ethico-political implications of archival uncertainty for the organization of knowledge today.

Kristin Veel is Associate Professor in the Department of Arts and Cultural Studies, University of Copenhagen. She is the Principal Investigator on the Uncertain Archives project, funded by the Danish Council for Independent Research.

  • Interview
  • 21 July 2017