The turn to ontology has established at least one indispensable insight: it has called attention to the fact that entities are intra-relational as well as inter-relational. It has compelled us to recognize that entities are intrinsically multiple, or self-differing. Without retreating from this insight, my contribution to this discussion will be to question whether intrinsic multiplicity necessarily implies an ontological—and therefore political—asymmetry between relations and entities. It has become axiomatic in some quarters that relations are logically prior to and encompass entities (e.g., Holbraad, Pedersen, and Viveiros de Castro 2013; Pedersen 2012; Viveiros de Castro 2010). The fact that entities comprise relations has been taken to mean that there can be no simultaneously autonomous things. Intrinsic multiplicity is presumed to constitute an invisible extensive pre-connectivity. But this asymmetry, I want to suggest, is not only unwarranted; it may also be politically undesirable.

To illustrate my point, I ask you to picture the image of Indra’s net, as developed in Chinese Buddhism. As many of you will know, Roy Wagner (2001) has invoked this image as an aid to conceptualizing what he calls the “holographic worldview.” Wagner tells us that the negative spaces—the holes in Indra’s net—are not really empty at all, but are “gems that reflect one another so perfectly ‘that they do not know whether they are one or many’” (2001, 13, quotation unattributed).

Wagner (2001, 13) suggests that this image instantiates what he calls “the absolute identity of part and whole.” His use of this image looks, in other words, like an example of what Eduardo Viveiros de Castro (2010) describes as a virtual connection between Wagner’s thought and the philosophy of Deleuze. Indra’s net is Wagner’s way of expressing what Deleuze and Guattari (2004, 23) call “the magic formula”: “PLURALISM = MONISM.” In both cases, the ontology indexed is one of infinite invariant fractality, what Wagner (1991, 163, 166) elsewhere describes as the “whole cloth of universal congruence” or “integral relationship” replicated across all scales. Everything contains everything else, at least in potentia.

Now, in my view, there are many potential problems with this holographic ontology, at least as methodological presupposition. For one thing, there is, at present, no evidence that the universe is comprehensively fractal—let alone fractal to the degree of invariant self-similarity across all scales. I am concerned that we have simply been wonderstruck by the apparent congruence between a few aesthetically powerful examples of invariant fractality—as described by scientists and mathematicians—and the familiar macrocosm–microcosm correlations found in many ancient, indigenous, and alter-modern cosmologies.

But the main point I want to make is this: if, like many of the ancient, indigenous, and alter-modern cosmologies we study, we posit an asymmetry between an all-pervasive relational background (whatever we call it) and entities, conceptualized as figures emerging from it, we risk reinventing—or lending support to—claims that some entities are either closer to, or somehow have greater access to their inner capacities for infinite becoming than others. Accordingly, if we return to the image of Indra’s net—as good to think if not to embrace as methodological ontology—we must acknowledge its absolute ambiguity. It is a classic figure–ground composition, but one that must be read alternatelyas either a radical or a partial duality (con. Viveiros de Castro 2010). It cannot be both at once only; a both/and formulation alone gives permanent ontological ascendancy to the “whole cloth” of relations over entities.

Wagner says that the gems do not know whether they are one or many. But it is equally the case, I suggest, that they do not know whether they are entities or relations. They do not know whether they are autonomous terms with their own core intra-relational essences, or nothing but nexuses in an infinite web. After all, if the negative spaces—the holes in Indra’s net—can be seen as positive, it is equally the case that the positive spaces—the ligatures—can be seen as negative, as gaps between the gems, rather than links. Indra’s net can instantiate a thoroughly essentialist ontology—one that posits autonomous multiplicities at every scale.

More importantly, intrinsic multiplicity—whether this means internal relations that are isomorphic at every scale, or (what is more likely) internal relations that are contingent and unique to every entity—need not preclude a priori autonomy. The insight that entities are composed of relations does not necessitate the asymmetrical privileging of relations over entities. People can—and, indeed, some people do—see the gems first as independent complexities in need of swerve, in need of external connections to start up a cosmos. At the very least, then, such a privileging of intrinsically multiple yet always already autonomous entities needs to be sustained indefinitely as a possibility, both in anthropological theory and in ethnographic contexts.