“Is there an ontology to the digital?”, Hannah Knox and Antonia Walford provocatively ask. Departing from this provocation, I offer some preliminary reflections on possibilities for intervening into the recent hype around big data by drawing on an interdisciplinary computational social science research project that I have been part of since 2013.

First of all, note the problematic tendency in ongoing discussions about the ontological turn to conflate an ethnographic focus on difference and alterity, as opposed to similarity and identity, with a metaphysical penchant for the many, as opposed to the one (Laidlaw 2012; Pedersen 2012). This misunderstanding seems to hinge on a deeper conflation of the contrast between extensive and intensive differences (see Deleuze 1994; De Landa 2002). For the moment that relations are conceived of as intensive as opposed to extensive, which is what the “methodological monism” of the ontological turn is all about, then the distinction between the one and the many—and derived binaries between individual and society, or one culture as opposed to another—ceases to make sense. Instead, the ontological turn's object of study is infinitely folded inward and outward, suspended between the radical contingencies of fieldwork and the radical reflexivity of the anthropologist (Holbraad and Pedersen, forthcoming). Far from choosing between the one or the many (humanism vs. posthumanism, universalism vs. pluralism, and so on), the binary between the one and the many is substituted by manifold partial connections (Strathern 2004). Which is why, as Knox and Walford remind us, the differences and the alterities at stake in the ontological turn are not so much between things as within them (Holbraad, Pedersen, and Viveiros de Castro 2014).

Accordingly, what I took with me from the digital ontology workshops in London was not an answer to whether the digital has one or more ontologies—another riff on the one/many question best left unanswered. Nor am I convinced that digital phenomena are distinct in having “the inherent capacity . . . to be continuously other than they are,” as Knox and Walford suggest. Other things have been shown to be imbued with such self-scaling properties (see Wagner 1991), including Mongolian shamanic spirits (Pedersen 2011) and Polynesian kinship-property systems (Miyazaki 2004). Rather, what we need to investigate is the precise manner in which digital phenomena are self-scaling in comparison to other self-scaling phenomena. Thus understood, the potential alterity of digital phenomena would not inhere in their ability to differ from themselves—indeed, all things are potentially self-different from the heuristic vantage of the ontological turn—but in the distinct way in which they could alter from themselves: the capacity of the digital, that is, to be differently self-different from other inherently self-different things.

This is the lesson I heeded from London: there is a need for a robust anthropology of the digital, but for this to come about we need to break free from the qualitative cocoon that has for so long defined our self-identity. In fact, what interests me about the digital ontology question is not so much the conceptual creativity that posing it unleashes, but the methodological innovations required to begin answering it. I am not suggesting that anthropologists, to explore the digital, renounce their commitment to qualitative research or their skepticism toward hegemonic quantification regimes. But there may be ways of intervening in such regimes that represent a useful complement to the emerging paradigm of critical big data studies among sociologists and anthropologists (e.g., boyd and Crawford 2012; Wang 2013; Stoller 2013).

To be sure, the time is ripe for big data bashing, judging from the arrogance of those scholars who consider classic social science to be on the verge of extinction (Anderson 2008; Lazer et al. 2009). Nevertheless, as I have argued with Anders Blok (Blok and Pedersen 2014), it is pressing to develop an alternative, more experimental ethnography in/of big data that is not less interventionist, but differently so. For could there not be important critical traction in actually doing big data studies by collaborating with computational social scientists, instead of treating them merely as objects of ethnographic description?

Such an approach might allow for a more direct understanding of the new digital ontologies emerging from big data experiments both inside and outside the academy. It might present possibilities for transforming the morphology, conceptualization, and visualization of the resulting datasets and method assemblages, which have been described as quali-quantitative (Latour et al. 2012). Arguably, this is what our team of anthropologists and sociologists has attempted to do through our participation in the interdisciplinary Social Fabric/SensibleDTU computational social science project (see Stopczynski et al. 2014; Blok and Pedersen 2014; Blok et al., in preparation). We have been trying to create the conditions of possibillty for an experimental ethnographic encounter that entails not a passive critique of big data, but active participation in the creation of partly connected data stitchings and critical algorithms.