by Thao Phan and Scott Wark
In the commentary ‘Racial Formations as Data Formations,’ we outline what we see as the challenges and implications for studying big data’s effect on the category of race itself. The commentary is a part of Big Data & Society's special theme, ‘Data, Power and Racial Formations.’
This essay is framed around a proposition made by Paul Gilroy in some of his earlier work on the link between race and technology. In the 90’s, Gilroy made the provocative argument that new scientific and medical techniques for imaging bodies ought to allow us to renounce the concept of race, because at the micro-scales imaged by, for instance, MRI scans, all bodies are the same.
Without necessarily endorsing this claim – old forms of racism have an unfortunate habit of persisting – we use Gilroy’s provocation to suggest that scholars need to pay more attention to the link between racialisation and mediation. For Gilroy, the category of race itself has to be understood as an effect of the techniques we use to represent it.
Big data provides a new means of mediating race visually – for example, through techniques like facial recognition. But it also operates using a post-visual logic. Large-scale data collection is post-visual because it happens at scales and through processes which we can not only not see, but which also produce effects that are not at all ‘seeable.’ It occurs in what we call a ‘post-visual regime.’
Whereas much of the – excellent and inspiring – existing research on data emphasises how computational systems reproduce historical inequalities by using biased data, we argue that scholars also need to account for their post-visual effects. Algorithmic techniques transform the inputs they work on. With enough data, for instance, people can be racialised using ‘proxy’ indicators of race, like taste, interests, or location.
Large-scale data processing doesn’t just reproduce existing racial formations, to use Michael Omi and Howard Winant’s hugely-influential term. It produces new ones. This is what we call racial formations as data formations.
This is our essay’s provocation: recapitulated as data formations, racial formations reconceive us in new ways that we cannot see and that we might struggle to access. By transforming the category of race itself, these formations also change the terms in which it might be contested.
