Work on the development of two relational database systems required us to confront assumptions about digital objects and the kinds of relations through which such systems might be constituted. In particular, it raised practical problems regarding database ontologies and how they might accommodate and allow exchange across different relational modes. Addressing these challenges raised questions about distinctions between real and virtual objects as well as between relativist and relational approaches. These kinds of differences are also being addressed in current anthropological discussions of ontology.

Whereas ontologies in information science are taxonomic hierarchies allowing data to be shared across diverse systems and platforms, anthropologists borrow the term from philosophy’s distinction between matters of being and of knowledge. In recent debates, ontology has acquired a variety of meanings, signifying anything from peoples’ theories of the world to subconscious assumptions about the nature of reality to the reflexively produced artifacts of scholarly analysis (Salmond 2014). These discussions are haunted by a specter of relativism, an association linked to the idea of mind as a ghost in the machine of matter (see Ryle 1949). Drawing on the legacy of such dualisms, critics have charged the ontological turn with advocating an extreme form of cultural relativism, in which peoples are seen to inhabit different worlds irrevocably divided by incommensurable concepts. Parallels are drawn with earlier debates on linguistic relativism (the idea that different languages entail different thought-worlds) and with theories on the relativity of scientific knowledge. Yet such comparisons often miss the emphatic post-representationalism of many current ontological projects.

Recursive ethnography, for instance, rejects the dualist thesis as an exhaustive account of reality, proposing instead—for heuristic purposes—to treat concepts and things as an identity (Henare, Holbraad, and Wastell 2007; Paleček and Risjord 2012). Its aim is not to advance a better universal ontology, but to generate methodological openings that might admit differences other than the ones anthropology already anticipates. Recursive approaches advocate ethnography as the method through which these other forms of alterity might emerge on the basis of its generative relationality, its production of insights through comparisons. In recursive terms, ontological difference is less a property of mental, linguistic, or cultural schemes “out there” which condition human apprehension of material reality than it is an artifact of comparative relations. The aim of recursive arguments is thus not to establish ontological alterity as empirical fact: by mapping divergent assumptions about reality onto groups of people, for instance. Rather, the ambition is to advance a methodology genuinely open to the unexpected and the otherwise—one that precisely refuses to place a bet either way when it comes to the question “what is?” (Pedersen 2012).

Analogously, our work on relational database systems required us to make comparisons between potentially different ways of relating and between unanticipated kinds of digital objects. We have collaborated on two systems: KIWA, a digital research environment built for researchers studying ethnographic museum collections; and Te Rauata, a digital taonga repository built by and for Te Aitanga a Hauiti, a Māori tribe from New Zealand. These database systems are designed to be partially connected, so as to exchange information about museum artifacts that originated in Hauiti’s tribal territories and are now held in overseas museums.

Initially, we thought KIWA and Te Rauata might take the form of a single digital research space that would accommodate diverse artifacts of knowledge (Hogsden and Poulter 2012a). It soon became clear, however, that KIWA’s artifact-centric structure would not work for Hauiti. Through a series of workshops and discussions, it emerged that Te Rauata was envisaged as an entirely separate system, with a structure and ontology that would enable Hauiti to maintain mana and control over their whakapapa, the generative and emergent constellation of relations that are both the substance and conditions of possibility for Hauiti’s ongoing existence (Salmond 2013). In whakapapa, the a priori condition of everything, including every object in the database, is its enmeshment in the relations that make it what it is. Nothing—no digital object—that was not already part of a system-internal relationship could enter Te Rauata. Both database systems are relational, but differently so. Whereas KIWA’s objects are connected by relations, Te Rauata is relation-centric: nothing inside it is not already related (Hogsden and Poulter 2012b).

In Te Rauata, digital objects relating (to) images, events, art forms, artifacts and language are valued in their own right, not least for their capacity to help maintain and restore the relational perspectives (whakapapa) that transform potential taonga into actual ones (virtual or otherwise). Hauiti’s concept of digital taonga thus underscores the relational nature of taonga in general (Ngata, Ngata-Gibson, and Salmond 2012), to the point that—in certain ways—distinctions between real and virtual forms disappear. Far from mere ghosts in the machine standing in for real material things, such objects stand on their own in relation to others, as the substance and conditions of possibility for both the dynamic unfolding of whakapapa, and for new digital ontologies.