In this essay, I explore how scientific data presents us with a form of temporary possession that calls forth a particular sort of subjectivity. There has been much debate about the prospect of digital media sidestepping older, proprietary configurations and constraints (as with Stewart Brand’s infamous claim that “information wants to be free” and the Free and Open Source Software movement; see Boyle 1996; Lessig 1999, 2008; Kelty 2008). This prospect depends, in part, on what are seen to be properties of digital culture: the ease of digital media’s copy-ability, its apparently frictionless transmissibility, and its capacity to be remixed, for example. But the ease of dissemination and the possibility of replication also depend on social and material contexts and relations. There is no doubt that digital forms have material properties (Dourish 2017), but these properties are never not in relation with other entities, forces, and social forms.

In the case of data, I would like to draw out one particular property that is important for questions around ownership. This is the way in which data is constituted by its potential: what Marilyn Strathern (1996, 17) has pointed to in a different context as a “capacity for development as yet unrealized.” Taking my lead from Strathern, I suggest that data is in some aspects very close to other forms of potential property that are difficult to recognize: embryos, seeds, ideas. Data is often coveted because it is understood as something that will transform into something else (see Gitelman 2013). Although this property applies to data as an emerging object of social interest across various domains, here I focus on how this plays out for the object of my ethnographic research: scientific data from the environmental sciences. The context of scientific practice gives the setting some specific contours; I do not have space here to go into the details of the place I conducted fieldwork in—the Brazilian Amazon—but clearly this is also important.

Property in science is a tricky business. According to Mario Biagioli (1998, 2003, 2008), scientific knowledge cannot be owned at all: scientists produce truth, and truth—as a representation of Nature—by definition cannot be anyone’s property. Additionally, the way scientists produce scientific knowledge also prevents them from owning it: the scientist can only discover and cannot be seen to create in any way, as this would disqualify their product from being true (Biagioli 2003, 257).

However, in the case of my fieldwork with environmental scientists and technicians working in a large scientific program in the Brazilian Amazon, things were more complicated. Scientific data is not truth, nor does it quite play the part of Nature. And, at least in my field site, it certainly elicits relations of ownership, circulating through and constituting various regimes of value. But it is its role in coconstructing scientific subjectivities to which I want to draw attention here.

An enlightening example is the case of the people who clean or curate data. In the project I conducted research inside, these data curators spend their time downloading data from research projects, processing that data, and making it ready to use by other researchers. They are paid to do so, and so money is used to sever their relationship to the data they produce (Strathern 2012). But the situation is not as clear-cut as this. The data curators did feel like the data was theirs. Beyond the liberal equation of their labor with ownership, there was also an embodied relation that they developed to the data. They spent hours with it, carefully teasing out its errors and flaws, standardizing it so that it was available for use by others. I was told that they had to learn how to “woo” (namorar) the data, to get to know it intimately. This was, however, tempered by a sense that the data was nonetheless not theirs, that the work they did was just a job. Caught between these two poles, the data cleaners quietly held a lot of frustration. Even though many of them held advanced degrees of their own, they sometimes lamented that they had no time to publish with the data they cared for. Without publishing, they felt that they had little means of developing as scientists. This left them feeling liminal and precarious.

I suggest that the position of these curators can be understood as one in which they own the data because of their work on it, but do not own the potential the data has in it. The data is split for them in a way it is not for other researchers, who collect data and publish with it, transforming it into knowledge. By contrast, the work the data curators do bestows a relation of ownership that they subsequently cannot turn into future relations that might constitute them as scientific subjects (cf. Strathern 1996, 18).

If we were to look for a comparison, it might be useful to think of the egg donors in Monica Konrad's (1998, 2005) research, who give away potential without any claim to its realization. Here too, “what signifies is being the origin of a process that another carries forward” (Strathern 2003, 185). What surprises in reading about these egg donors is their capacity to be anonymous and nonpossessive in relation to their ova. In the case of data curators, they are not expected or permitted to experience a sense of ownership, but they nevertheless do. They are caught up in an ongoing, but also strangely truncated relation with the data. This partial possession is almost the opposite of the conventional understanding of usufruct: they have the right to possess but not to use it; the right to feel ownership toward it but not transform it. This eclipsed possessiveness is also why the data cleaners feel liminal and stuck. Open and unrestricted movement, endless copy-ability and remixability do not make much sense in this context; they belong to a different imaginary of property and a different set of properties.