BY 4.0 license Open Access Published by De Gruyter Open Access May 27, 2020

Virtuality and the Problem of Agency in Object-Oriented Ontology

Ruslanas Baranovas
From the journal Open Philosophy


In his Prince of Networks, Graham Harman reconstructs Latourian critique of concepts of potentiality and virtuality with which he claims to agree. This seems striking because Latour’s arguments seem to be exactly those Harman rejects in his other writings as overmining. Furthermore, this critique of potentiality and virtuality creates a dividing line between Harman and Bryant’s Democracy of Objects, where the concept of virtual plays a central role. In this article, I will explore this debate, focusing on how the concept of virtuality works in the context of the ontological realism that Object-Oriented Ontology is. To do this, I will first present Bryant’s notion of virtuality focusing on the problem of the individuality of the object. Then I will explore Latourian–Harmanian arguments against virtuality and show that the main issue Harman has with virtuality has to do with the agency of objects. Therefore, I claim that the main dividing line between Bryant’s and Harman’s versions of Object-Oriented Ontology is the difference between the two notions of agency.

Under the spell of Bergson and Deleuze, the critique of the classical notion of potentiality in the name of some sort of notion of virtuality became a big trend in contemporary continental philosophy. After Deleuze, the notion of virtuality played a crucial role in many of the philosophies of his followers, most notably in DeLanda’s Intensive Science and Virtual Philosophy and ignited fruitful debates with other traditions. For example, it is in the discussion with this tradition that Agamben developed his notions of potentiality-not-to and Potenza. In this article, I will explore what happens when this Deleuzian notion of virtuality is transplanted into the speculative realist soil. More specifically, I will focus on the debate on virtuality in the context of Object-Oriented Ontology, with two key figures in this debate being Graham Harman and Levi Bryant.

1 The briefest possible introduction to Object-Oriented Ontology

Let us start with the briefest possible introduction to Object-Oriented Ontology (I will further use the abbreviation OOO). I will confine myself to two theses that, I think, capture the essence of OOO: first, only objects exist. This means two things. First, objects are individuals who can come into relations. In Harman’s original version of OOO, objects are usually considered from the perspective of an encounter between objects, rather than some theoretical speculation about the objects outside of the encounters among them. In this sense, Harman’s OOO is much more phenomenological than other more theoretical versions of speculative realism (take, for example, Meillassoux perspective of hyperchaos). It is crucial to note that what qualifies something as an object is not its size or ontological proximity to our everyday experience. Rather, an object is a self-standing source of determination. OOO would treat a chair, sunglasses, mountain, atoms, planets, the universe, Bruno Latour, or a plane crash as objects. Second, the thesis that only objects exist means that everything that is is an object. Therefore, all objects are ontologically equal, meaning that they all are self-standing sources of determination. Therefore, OOO is sometimes labeled as flat ontology. The second thesis, with which I will complete this introduction, is the following: No object can know or encounter other object-in-itself, but only its appearance. The real object always withdraws. Harman puts it this way: “Just as human cognition of a cotton ball is finite, governed by space, time, and the twelve categories, so too is the fire-cotton interaction a drastic oversimplification of these two entities by one another. […] I encounter the cotton ball in one way, the fire encounters it in another, the boll weevil in still another”. [1] This means that while objects are the sources of determination in the world, so are other objects. And the true encounters between the objects always entail a certain entanglement of the two sources of determination. Another way to put this would be to say that not only does OOO not return to some version of “myth of the given” [2], but rather claims that there is no “pure given” in any encounter between any two objects.

Harman’s technical term for this direct inaccessibility of an object is “withdrawal”. According to the theory, in every encounter, an object splits into a sensual object, which is manifest in the given encounter, and the real object, which always withdraws from it. According to Harman, “OOO’s commitment to the mutual darkness of the objects is what enables it to resist some of the fashionable holistic philosophies of our time”. This is the case, because the withdrawal of the real object ensures that it is always more than its sensual manifestation, which consists of relations. [3]It seems to me that the sensual object is the result of the indirect entanglement of two sources of determination qua real objects. However, this does not result in “a two worlds theory”, as there is no hidden dark world of real objects. Rather, one could speak of the countless spaces occupied by the real objects. As Wolfendale notes rather poetically, “What we have here is a pluriverse of infernal engines that present themselves to one another only so as to hide their internal machinations. Like the many hells of Buddhist lore. each of these engines is a realm unto itself”. [4]

2 The paradox of substance and Bryant’s goal

Such an introduction instantly raises the questions: what can be known of these withdrawn objects, given what they are withdrawn from our epistemic (and other) capacities? Are not such objects, by definition, left completely underdetermined, or vacuous, as both critics and Harman claim? In his Democracy of Objects, Levi Bryant sets out to answer these worries. He formulates the problem of the withdrawn object in the terms of “the paradox of the substance” (“substance” is one of the terms Bryant uses for the withdrawn object):

If substance is not its qualities, does this not entail that substance as such is without qualities and is therefore a bare substratum? And if substance is a bare substratum, does this not entail that all substances are identical? If this is the case, then this spells the ruin of the concept of substance for substance is supposed to account, in part, for the individuality of substance. Yet where substance is bare, all individuality is erased. If this difficulty is to be avoided, we require some way of talking about the structure or formatting of substances or split-objects without this structure consisting of qualities. [5]

We can reformulate the paradox of substance in two mutually exclusive, yet both unsatisfying, solutions: the substance cannot be determined by its qualities, because an object might undergo a change in determinations while remaining the same object. [6] One can avoid such charge by making a distinction between essential and accidental qualities. However, such move would face two substantial problems: first, a successful attempt to rigorously delimit the essential and accidental determinations for any given object is still to be produced (Plato’s dialogues alone are proof enough on this matter); second, the thinking of essences is usually deemed as opposed to historicity and the concrete labor of entities in the world (if not altogether condemned as totalitarian) and thus has fallen out of fashion in philosophical circles. But if we reject that the substance is “made” out of determinations of an object, we are forced to posit substance as a bare substratum, which seems to be incapable of individuating an object. [7] Therefore, we cannot give an account of either change or individuation. As I will shortly show, what first and foremost concerns Bryant in relation to OOO is the second question of how the withdrawn object can carry out the work of individuation, because the paradox, as Bryant calls it, is also an internal critique of Harman’s withdrawn object, inasmuch as the argument on the individuation could equally be applied to it.

Bryant thinks that the stage is now set for his own solution, which lies in thinking “about the structure or formatting of substances or split-objects without this structure consisting of qualities”. [8] At this stage, the concept of virtuality will take the central stage in both Democracy of Objects and Bryant’s debate with Harman. Therefore, before proceeding to Bryant’s notion of virtuality, it is important to recollect all that are being thought under this concept. First of all, virtuality must help us think how the new and novel enter the world. Second, virtuality must also provide the means to individuate the real object and think it in a more concrete way. Third, all these must be done in the realist manner, that is to say, virtual must in some sense belong to the object and not vice versa.

3 Bryant’s notion of virtual proper being

Bryant’s notion of the virtual could be summarized in four points.

3.1 Reinterpretation of Deleuze

Bryant’s notion of virtuality is mostly built on the reinterpretation of the same notion in Deleuze. Symptomatically, Bryant starts the chapter on virtuality by quoting Deleuze: “In Difference and Repetition, Deleuze remarks that ‘[t]he virtual is opposed not to the real but to the actual. The virtual is fully real in so far as it is virtual. Exactly what Proust said of states of resonance must be said of the virtual: ‘Real without being actual, ideal without being abstract’”. [9] However, to understand what truly inspires Bryant, one has to keep in mind Deleuze’s idea that the ground or the cause is radically different from its effect. In this context, this means that the virtual does not resemble the actual. To illustrate this point, Bryant appeals to the example of genes. Genes structure the phenotype of an organism without in any way resembling the phenotype itself. Genes themselves do not have eye color or eyes or color for that matter. According to Bryant, “this captures the manner in which the virtual dimension of a substance differs from anything qualitative, thus providing us with substance that is structured or formatted without being qualitative”. [10] This helps him rethink Harman’s distinction between real and sensual objects. In Bryant’s version of OOO, the real object turns into virtual proper being (powers and structure of the substance) and sensual object becomes local manifestation (qualities and extensities).

Bryant also follows Deleuze in ascribing the relational nature to the virtual. According to him, “In claiming that the virtual is ‘ideal’, Deleuze is not claiming that the virtual is mental or cognitive – though minds too have their virtual dimension – but rather that the virtual is relational. These relations, however, are not relations between entities, but constitute the endo-structure of an object, its internal topology”. [11] In other words, the unity of substance itself gets differentiated and relational. For Bryant, substances are not rocks, but lava under the appearances. In this way, Harman’s real and sensual qualities are changed into endo (inner) and exo (outer) relations. I will say more about this later, but it shows that Bryant treats substances as organizations of powers. According to him, such treatment of substances closely follows the Deleuzian notion of multiplicity. [12]

3.2 Withdrawn virtuality

Exactly like Harman’s withdrawn object, the virtual proper being of an object is “perpetually withdrawn” from the plane of local manifestations. As Bryant puts it, “no one nor any other thing ever encounters an object qua its virtual proper being, for the substance of an object is perpetually withdrawn or in excess of any of its manifestations. Rather, the virtual proper being of an object can only ever be inferred from its local manifestations in the world”. [13] This is probably one of the most defining characteristics of Bryant’s notion of virtuality. As he puts it, “the virtual is entirely real without, for all that, being actual”. [14] The virtual never becomes fully actual but always remains a withdrawn excess from which the change comes. What is more, withdrawn virtuality is the source of change and identity of the object, without being inscribed in it in advance, as potentiality allegedly is. This change in terminology also helps Bryant to stress two points which are already present in Harman but could be obscured by his terminology. On the one hand, the virtual puts a bigger emphasis on the withdrawn object as the locus of change than Harman’s real object does. On the other hand, “local manifestation” seems to avoid needless subjectivist connotations that the “sensual object” has.

3.3 Belonging

It is necessary for Bryant to bind the virtual to the individuality of an object in order not to end up with some postmodern sounding “virtual as pure difference” type of notion. On the contrary, “The virtual proper being of an object is what makes an object properly an object. It is that which constitutes an object as a difference engine or generative mechanism” and “The virtual proper being of an object is […] being as a (more or less) enduring unity”. [15] To understand this move, one has to keep in mind that Bryant’s onticology and OOO in general are ontological realisms about individuals. Put differently, here the individuals are taken to be the primary source of determination in the world (be they objects or subjects, that does not matter). Now if we think of the individuality of the object itself in terms of the virtual, the virtual itself must be individual through and through because, in the other way, the objects that are supposed to be primary would be constituted through something else. Perhaps, there is no virtual as such then, only virtuals. However, the problem arises here because if, as Deleuze maintains, the virtual ground in no way resembles the actual, shouldn’t this mean that the virtual cannot be individual because the actual objects are? Harman will contest these two last points: the individuality and belonging of the virtual to a certain topos.

3.4 Powers

Finally, such a notion of virtual proper being allows Bryant to think of objects dynamically. As he puts it, “the substantiality of objects is not a bare substratum, but rather an absolutely individual system or organization of powers. Powers are the capacities of an object or what it can do”. [16] Again, keep in mind that the OOO type of ontological realism claims that it is individuals who are the source of determination in the world. I think the language of powers resonates beautifully with such realism. Consider Bryant’s remark on qualities as powers when he claims that “rather, if we had an ontologically accurate language, we would instead say that ‘the mug blues’ or that “the mug is bluing” or that “the mug does blue […] The blueness of the mug is not a quality that the mug has but is something that the mug does”. [17] At its best, Democracy of Objects gives a very strong sense of agency to objects.

The question is, however, why not also claim that “the object mugs” or that “the object is muging”? In other words, one could develop Bryant’s point to a claim that even concrete individuality of an object is also only one of its powers. Doubtlessly, Bryant himself would reject such a conclusion, as it would postulate something like Leibniz’s monads behind all change in the local manifestations. Remember that one of the primary aims of Democracy of Objects is precisely to give more concreteness and efficacy to Harman’s withdrawn object. Therefore, a move to something like the monad model would be contrary to the intention of the book. Nonetheless, such traction toward “monadism” certainly exists in the ranks of OOO.

To summarize, for Bryant the virtual is something that structures the being and unleashes the powers of the withdrawn object. Doing so, the virtual is capable of individuating objects and unfolding their dynamic being. However, I do not think that Bryant would agree if one would equate the two and claim that “the withdrawn object is virtual”. This would imply that the substances are not actual. And while it makes sense to claim that “the substance as withdrawn object” is not fully actual in relation to the local manifestation, it is hard to see how one can think substance in itself as fully virtual, and therefore, not actual. Therefore, the substance = virtual equation does not hold. I think that at minimum, some part of the endo structure of the object must be actual.

4 The unexpectedness of the critique

The idea that we should conceive the withdrawn object as excess with regard to phenomenal and the source of novelty in the world is central in Harman’s version of OOO. For this reason, Harman’s “vigorous critique” [18] of the potential and the virtual seems to be unexpected and strange. This also seems to be the feeling Bryant has, who at least on two occasions tries to reaffirm the similarity of his position to Harman’s (for example, Bryant states: “I sense that my position and Harman’s are much closer than might initially be suggested”). [19] Given all the similarities between the positions of Harman and Bryant, one may even suggest that the differences between them are mostly terminological. However, Harman’s unexpected vigorous attack on all notions of virtuality that stirred the debate with Bryant signals that this is not the case. I hope that this article is successful in showing what philosophical differences between Bryant and Harman may be.

Another source of difficulty with Harman’s critique of virtuality is that it is done indirectly. Harman develops his critique of virtuality in his book on Bruno Latour, Prince of Networks, which, according to Harman “has described Latour as a pioneer of object-oriented philosophy”. [20] However, even if Harman’s critique is certainly mediated through Latour’s “actor-network theory”, he himself claims to subscribe to the critique when he states that “on this point I can only salute Latour”. [21] However, this seems to be even more strange, as the claims which he salutes Latour in seem to be exactly the same claims he rejects in his other works. For example, in Prince of Networks, he takes the side of Latour and ancient Megarians over Aristotle in the debate on the so-called actualism: “To speak of something existing in potentia implies that it is already there but simply covered or suppressed. This is what Latour denies. […] For Aristotle, a thing is always more than what it is right now; for Latour, never”. [22] But this seems to directly contradict the fact that Harman constantly takes the side of Aristotle over Megarians in his other works. Take, for example, his argument against overmining in the latest Object-Oriented Ontology: New theory of everything: “If atoms, billiard balls, watermelons, prisons or Professor Whitehead are nothing more than the sum total of their relations or effects in this very instant, then how is it that they can be doing something very different five minutes or two weeks from now? Aristotle raised this very question […] against […] the Megarian school of philosophy”. [23] So, on the one hand, Harman seems to salute Latour, who claims that “a thing is never more than what it is right now”, and on the other, he seems to admire Aristotle who denies that things are “the sum total of their relations or effects in this very instant” (emphasis is mine).

In what follows, I will try to show that these contradictions are only apparent. In fact, Prince of Networks presents a different argument than that directed against overmining. Contrary, I will try to show that the same motivation is behind both arguments against actualism and actualist argument against potentiality. But for this, I need a closer analysis of the Harman–Latour argument and Bryant’s response to it.

5 Harman and Latour against the potential and the virtual

There are two principal points intertwined in the Harman–Latour critique of the virtual and the potential. But first, note that under the Latourian scheme, as I will show, actualism is opposed to both potentiality and virtuality, and these two obviously cannot be equated in all aspects. On the contrary, the concept of the virtual itself was first developed to overcome problems philosophers had with the concept of potentiality. Therefore, Harman’s first argument may seem to be directed exclusively against potentiality, but as I will argue, that is not the case. Let us return to the quote already mentioned in this article. With a pretty traditional observation-critique, Harman claims that “to speak of something existing in potentia implies that it is already there but simply covered or suppressed”. [24] In other words, according to the defenders of potentiality, things always have their possible futures inscribed in them in advance. The concept of virtual was developed precisely against this teleological predetermination, especially in the thought of Deleuze. In fact, Bryant provides a remarkably similar critique of the potentiality to that of Harman’s: “The great error to be avoided lies in conceiving the virtual or potential in teleological terms”. [25] Therefore, from Bryant’s point of view, it seems that the Harman–Latour critique does not actually apply to his notion of the virtual. According to him, “in this regard, Harman and Latour seem to conflate the virtual with the possible”. [26] However, I think that the point is not so much in the critique Harman–Latour provides but in their alternative solution to the problem. According to Harman–Latour:

To make something become actual is not to unfold a cryptic seed lying hidden in the ground, but to assemble a wide range of actors that begin in separation. For instance, it would make no sense to call Joliot “the potential father of the atomic bomb”, except in a purely literary way. What Joliot would have needed to do, and what Oppenheimer did do, is help piece together an assemblage of neutrons, heavy water, uranium, cannon assemblies, plutonium, metallurgists, security guards, loyalty oaths, government funds, and so on. [27]

In this sense, Latour is an actualist: there is no potential atomic bomb before all these agents enter into alliances creating one. According to such a stance, the change in the world is produced by the concrete labor of objects, rather than the actualization of something inscribed in them in advance, or, and this is the point, befallen on them by chance. Note that, crucially, it would not be much better if we would change the potential for the virtual here as all the same difficulties would still apply if Bryant would claim that “the atomic bomb is the result of the virtual powers of all objects which entered the assemblage called the atomic bomb”, or that “before the creation of the bomb Oppenheimer had the virtual powers to create it”. Therefore, from the actualist point of view, there is no difference between potentiality and virtuality on this issue. In other words, virtual still has to precede the actual both logically and chronologically for Bryant. He cannot turn to some retroactive notion of virtual ala Badiou or Žižek because then the problem of the agency of change would remain untouched.

Bryant rejects such an actualist stance. According to him, it implies that entities (such as “the atomic bomb”) jump into being out of nothing. [28] And he thinks that it is simply not how things stand: “Yet if this is the case, it is difficult to see how we can get from one entity to another entity. Rather, it seems that entities must possess the capacity, the potentiality, to undergo change”. [29]

However, I think that what is truly at issue in this debate reveals itself in Harman’s second objection against the virtual and potential. Bryant himself thinks that “the real motivating desire behind this [Harman’s – R.B.] critique is the desire to preserve the concreteness of objects”, [30] but I would argue, what the real issue for Harman is not so much the concreteness, as the agency of objects as the sources of determination. Before going to the second objection, it is critical to underline that the notion of agency here does not imply some version of panpsychism. That is to say, I do not want to imply that for Harman all objects have intentional consciousness, make decisions, etc. In this context, I propose to understand agency in the following way: an agent is an object that not only is determined by other objects but also determines something out of itself, in his own right. Keeping this in mind, let us get to the second objection. According to Harman,

The recourse to potentiality is a dodge that leaves actuality undetermined and finally uninteresting; it reduces what is currently actual to the transient costume of an emergent process across time, and makes the real work happen outside actuality itself. The same holds true if we replace “the potential” with “the virtual”, notwithstanding their differences. In both cases, concrete actors themselves are deemed insufficient for the labour of the world and are indentured to hidden overlords: whether they be potential, virtual, veiled, topological, fluxional, or any adjective that tries to escape from what is actually here right now. On this point I can only salute Latour, the ancient Megarians, and other defenders of actualism. [31]

While the first sentence of the quotation suggests that what primarily worries Harman is some version of overmining implied in the virtual and the potential, the second part of the quote concerns precisely the agency of objects. I think that Harman here tries to point out that virtuality (just like potentiality) places into the object another source of determination which by definition the object itself does not control but rather is being controlled and determined by. And while Bryant claims that “the process of actualization is a creative process within substances that requires work”, [32] this does not address the issue of agency mentioned above. And while he appeals to Deleuze, who states that “[t]he actualization of the virtual always takes place by difference, divergence or differentiation”, [33] this does not seem to help either, because the logic of differentiation does not put any more agency into the object itself (while it would be a valid answer if the charge was against teleology supposedly implied in the virtual). By definition, Deleuzian virtuality seems to be something that happens to the object, rather than something object could possess or initiate. Be it otherwise, it would be hard to see how one would distinguish it from actuality and how it could escape the problems Deleuze had with potentiality. If one would claim that in virtual proper being of Oppenheimer there were some “powers or attractors” (to use Bryant’s terminology) toward the atomic bomb, then (1) these would seem to be already actual is some sense and (2) one would have to multiply powers or attractors infinitely to account for all the things Oppenheimer could have done. Bryant could escape these problems claiming that the virtual is not the power to do (or not do) something but rather power as such. However, this would be as abstract as Harman’s excess which lies in the withdrawn object, so why bother with virtuality at all?

6 Agency and the critique of mining operations

I think that the problem of the agency is also the underlying motivation behind Harman’s critiques of undermining and overmining. These critiques may strike as general critiques of reductionism. More concretely, the reduction of objects we meet in our ordinary experience. If that would be all there is to it, this critique would simply amount to the expression of preference Harman has to objects we experience in our everyday lives. However, we can see that this is not the case when we observe that for Harman objects of all ontological scales (waves, atoms, rocks, humans, and Dutch East India Company) are equally objects. This means that all the objects are equally sources of determination even if some of them make more difference in the world than others. Therefore, the point is not that Harman protests against the reduction of things we encounter every day to some hidden forces to preserve the world we experience, but that undermining and overmining reduce objects to nonobjects, that is to say, only the expressions of true sources of determination.

In this sense, Latour is at the same time, an ally and a foe to Harman’s realism. The primacy of networks over objects diminishes objects as the sources of determination, but the protest against potentiality and virtuality tries to protect the very same agency of objects as sources of determination.

7 Two freedoms, two democracies

I claim that the two notions of agency and freedom are at stake in this discussion. For Bryant, the democracy of objects lies in their horizontality and equal ontological rights. He faithfully follows Ian Bogost’s thesis that “all objects equally exist, but not all objects exist equally”. [34] However, the life of objects is not fully in their control: objects are attracted to one or other direction by the powers of the virtual, they undergo constant mechanical differentiations which are outside their control, etc. In his later work, Ontocartography, [35] he abandons the term “object” and instead uses the term “machine”. However, the flatness and nonreflective character of objects, which machine captures so well, are already evident in the Democracy of Objects. One of the most intriguing aspects of Bryant’s theory is that he introduces self-relations (endo-relations in his terms) into the withdrawn object; however, these self-relations do not seem to introduce any reflective character into the object. [36] One could call Bryant’s position a certain version of compatibilism: objects are free to express and unleash their forces unto the world, but these forces are often determined by something that the object strictly speaking is not.

That is not to say that against this Harman defends some liberal theory of freedom, where objects are the absolute sources of agency in the world (sovereigns, to employ Bryant’s expression). However, some version of the theory of self-determination seems to be in the back of his head, when he worries that virtuality could turn into “the hidden overlord” for an object. The spontaneous fluctuations of the virtual are as debilitating to the agency of the object as a predetermined possibility. And I think that surely both democracy and ontological realism must presuppose self-determinacy and agency.


This article is part of the European research project “The Future of Humanity: New Scenarios of Imagination” (Vilnius University). This project has received funding from the European Social Fund (Project No. 09.3.3-LMT-K-712-01-0078) under a grant agreement with the Research Council of Lithuania (LMTLT).


Bogost, Ian. Alien Phenomenology, or What its Like to be a Thing. Minneapolis: University of Minnesota Press, 2012. Search in Google Scholar

Bryant, Levi R. The Democracy of Objects. London: Open Humanities Press, 2011. Search in Google Scholar

Bryant, Levi R. Onto-Cartography: An Ontology of Machines and Media. Edinburgh: Edinburgh University Press, 2014. Search in Google Scholar

Harman, Graham. Prince of Networks: Bruno Latour and Metaphysics. Melbourne: Repress, 2009. Search in Google Scholar

Harman, Graham. Bells and Whistles. More Speculative Realism. Winchester/Washington: Zero Books, 2013. Search in Google Scholar

Harman, Graham. Object-Oriented Ontology: A New Theory of Everything. London: Pelican, 2017. Search in Google Scholar

Wolfendale, Peter. Object-Oriented Philosophy: The Noumenon’s New Clothes. Falmouth: Urbanomic Media Ltd, 2014. Search in Google Scholar

Received: 2020-01-07
Revised: 2020-03-13
Accepted: 2020-04-09
Published Online: 2020-05-27

© 2020 Ruslanas Baranovas, published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.