Mapped,Measured,and Mined: The Social Graph and Colonial Visuality

The social graph is a visual representation of relationships between social media users, arranged as a web of networked nodes. As a relational map, the social graph guides the development of a socio-technical assemblage that facilitates major social media platforms’ fundamentally commercial organization of social life online. The purpose of constructing this kind of representation is to support sophisticated queries, including such programs as collaborative and social filtering, where algorithms render social interactions into traceable, measurable, and thus commodifiable relations within the graph.

One of the most striking examples of the social graph’s iteration in contemporary digital media is of course Facebook, helmed by chief executive Mark Zuckerberg. Facebook’s social graph has proven the effectiveness of the graph as an infrastructural parameter that automatically and immanently transforms the social into something measurable, filterable, and scalable (Gerlitz & Helmond, 2013).

But perhaps more ominously, the social graph also underlies the “counter terrorism” efforts of mass surveillance programs that use filtering to trace “high value targets.” Among the Snowden cache of leaked National Security Agency (NSA) documents, NSA slides show how the social graph is used to track such targets through their networks, predicting impending “terror attacks.” Yet, through programs like Prism, the NSA has access to vast troves of metadata—not only the information associated with targets or security threats.

NSA whistleblower Bill Binney (2013) has said that the agency’s goal is to “build a social network of everyone in the world.” Doesn’t this sound suspiciously like Zuckerberg’s goal? And moreover, doesn’t this military–industrial thirst for mapping and arranging populations in order to control and/or exploit them evoke a long history of cartography and colonial modes of control?

Consider the relocation and reconfiguration of othered bodies in contexts like plantations, war zones, and camps. These older forms of visual social arrangement characterize the power-geometry of the way that capitalist enterprises grew and expanded globally, and they continue to shape contemporary time–space compression through digital technology (Massey, 1993). Capitalist expansion relied on disciplining different bodies “with the aim of making better use of them” (Mbembe, 1992, p. 12), reflecting the underlying logic of power-geometry as one of containment but also measurement designed to create value. In this way, the social graph and its associated contemporary surveillance economies might be situated within a longstanding colonial tradition of harnessing visuality for control and profit.

Visuality is a Western social technique for ordering based on the modern primacy of vision (Foster, 1988). Visuality invokes a socialized and historicized vision, situated within particular scopic regimes, for example, the scopic regime of modernity (Crary, 1991; Jay, 1988): Here, the observer, the one with the “right to look,” holds the power. As Mirzoeff (2011) argues, the establishment of visuality as a mode of power in visual classification is embodied through the colonial institution of the plantation, and especially the figure of the overseer. In the historical trajectory of visuality’s deployment, the contemporary military–industrial phase transplants the role of the overseer into the realm of “big data.” Mapping is now driven by data points rather than physical territory, but it still seeks to control bodies through a process of classification, differentiation, and reconfiguration.

The social graph facilitates discriminatory segmentation or even a micro-segmentation into sub-individual units based on specific aspects of the data set (Terranova, 2004). Facebook employs graph-based filtering to recommend ads to specific targeted audiences, profiled through processes of measurement and data mining (Turow, 2011). There are also more hidden intermediaries and data brokers, like Axciom and Datalogix, that exemplify how the micro-segmented audience entails a differentiation in relation to not only information but also the user’s physical body in a synergy between anatomical and numerical control (Appadurai, 1990).

The issue of control becomes perhaps more salient in the case of the military uses of the industrially established social graph. The Obama Administration’s “big data initiative,” for example, enables state access to private and commercial metadata. And of course, revelations about the NSA’s Prism program show how commercial social media systems have built-in vulnerabilities that are used as backdoors for mass surveillance.

The visuality employed in the social graph can thus be located within the late-modern military–industrial scopic regime. While the impulse toward exercising an omniscient point of view that exerts visual power through the disciplinary and territorial deployment of mapping has a long history, networked information comes with its own flavor of automation and immanence. In the social graph, information itself acts as agent (Braman, 2009). In this sense, the social graph performs a kind of autonomous visuality, rendering bodies culturally, socially, and politically intelligible through the relationships between data points on a massive scale that enable the identification of difference, of “anomalous” behaviors (Willson, 2014).

While the social graph has technical limitations that mitigate its potential for systemically embedding power relations across social media, such technicalities point toward the ideological salience of the social graph’s naturalization as efficient and invisible infrastructure. As Mirzoeff (2006) contends, visuality can be used as “both a mode of representing imperial culture and a means of resisting it through reverse appropriation” (p. 54). Attempts to make the invisible visible support the creation of competing scopic regimes, where a distributed plurality of vision might hold the possibility of overcoming networked absolutism.