The Democracy of Objects

1.1. The Death of Ontology and the Rise of Correlationism

Our historical moment is characterized by a general distrust, even disdain, for the category of objects, ontology, and above all any variant of realism. Moreover, it is characterized by a primacy of epistemology over ontology. While it is indeed true that Heidegger, in Being and Time, attempted to resurrect ontology, this only took place through a profound transformation of the very meaning of ontology. [8] Ontology would no longer be the investigation of being qua being in all its variety and diversity regardless of whether humans exist, but rather would instead become an interrogation of Dasein's or human being's access to being. Ontology would become an investigation of being-for-Dasein, rather than an investigation of being as such. In conjunction with this transformation of ontology from an investigation of being as such into an investigation of being-for-humans, we have also everywhere witnessed a push to dissolve objects or primary substances in the acid of experience, intentionality, power, language, normativity, signs, events, relations, or processes. To defend the existence of objects is, within the framework of this line of thought, the height of naïveté for objects are held to be nothing more than surface-effects of something more fundamental such as the signifier, signs, power or activities of the mind. With Hume, for example, it is argued that objects are really nothing more than bundles of impressions or sensations linked together by associations and habits in the mind. Here there is no deeper fact of objects existing beyond these impressions and habits. Likewise, Lacan will tell us that “the universe is the flower of rhetoric”[9], treating the beings that populate the world as an effect of the signifier.

We can thus discern a shift in how ontology is understood and accompanying this shift the deployment of a universal acid that has come to dissolve the being of objects. The new ontology argues that we can only ever speak of being as it is for us. Depending on the philosophy in question, this “us” can be minds, lived bodies, language, signs, power, social structures, and so on. There are dozens of variations. The key point here is that it is argued that being can only be thought in terms of what Graham Harman has called our access to being.[10] As such, ontology becomes not an interrogation of being as such, but rather an interrogation of our access to being. The answer to the question, “what is being?” now, everywhere and always, carries a footnote, colophon, or bit of fine print such that the question must be read as “what is being for us?”

And if the question of ontology now becomes the question, “what is being for us?” it follows that there can be no question of what being might be as such, for we have resolved to treat being only in terms of our access to being such that what being might be apart from our access to being now becomes an entirely meaningless question. This for two reasons: First, were we capable of knowing being apart from our access to being, it is argued, it would follow that we therefore have access to this being, thereby converting this being alleged to be beyond its givenness to us back into being-for-us. Second, to know something, the argument runs, we must have access to that thing. Yet being beyond our access to it is precisely a form of being to which we have no access. Therefore it follows that claims about such a being are, strictly speaking, meaningless. I hope to show later why there is good reason to doubt the soundness of both of these arguments, but for the moment it is enough simply to tarry with them to understand their logic, for these arguments constitute the basic schema of nearly every reigning philosophical position today.

If, then, these arguments are granted, it follows that there can be no question of what being might be like apart from our access to being. For the condition under which it would be possible to speak of being apart from our access to being would require access to that being, yet we do not have such access. Consequently, it follows that philosophy must abandon the question of whether being as it is given to us is like being as it is in-itself because we are unable to “get out of ourselves” to compare being as it manifests itself to us with being as it is in-itself apart from us. The best we can hope for, the most we can know, is being as it manifests itself to us, and philosophy shifts from being a discourse concerned with the being of beings, with what substance, as Meillassoux has put it, constitutes true substance (objects, God, nature, particles, processes, and so on?)[11], to being a discourse about the mechanisms through which beings are manifested to us (mind, language, normativity, signs, power, for example). Ontology becomes transcendental anthropology and the world becomes a mirror in which we don't recognize our own reflection.

Philosophy, of whatever stripe, thus comes to be characterized by what Meillassoux has aptly named correlationism. As Meillassoux puts it, “[b]y 'correlation' we mean the idea according to which we only ever have access to the correlation between thinking and being, and never either term considered apart from the other”.[12] While I have reason to disagree with Meillassoux's proposal for escaping the correlationist circle (and with his ontology), I believe that his concept of correlationism nicely summarizes the episteme, in the Foucauldian sense, that governs contemporary philosophy. When, in The Order of Things[13] and The Archeology of Knowledge [14] , Foucault introduces the concept of epistemes, he intends epistemes not as determinate positions in some discipline that might be opposed to another position, but rather as a set of statements functioning as the historical a priori of a particular discursive space. An episteme would be precisely that set of statements that allow opposed theories of, for example, language, to be opposed. It is the common framework these opposed positions share that allows them to enter into antagonistic relations with one another. The claim that correlationism constitutes the episteme of contemporary philosophy is thus not a claim about any specific philosophical position, but rather about the common framework regulating philosophical discourse in our contemporary moment. This episteme is shared by thinkers as diverse and opposed, for example, as the early Wittgenstein of the Tractatus, the Merleau-Ponty of The Phenomenology of Perception, and the Derrida of Of Grammatology. Despite the vast differences and disagreements among these positions, their thought and disagreements nonetheless unfold within the horizon of the unspoken premise of a necessary correlation between being and thought. The object of these disputes is not whether correlationism is true, but rather how the primordial or most fundamental correlation is to be articulated.

With correlationism we thus discover the root of contemporary theory's suspicion of both objects and realism. Realism must necessarily be anathema to all variants of correlationism by virtue of the fact that it claims knowledge of beings independent of the correlation between thinking and being. All realisms are committed to the thesis that it is possible to know something of beings independent of their being-for-thought, yet this is precisely what is precluded by the correlationist gesture. Here it is important to be precise. Correlationism is not the thesis of subjective idealisms whereby esse est percipi or where to be is to be perceived. Subjective and absolute idealism are only two variations of correlationism. The correlationist need not be committed to the thesis that there is no being apart from thought. Indeed, most correlationists are committed to the thesis that there is something other than thought. Kant, for example, held that in addition to phenomena (beings for-us) things-in-themselves exist. The correlationist merely argues that we have no access to these beings that are apart from thought and can therefore only speak of being as it is for-us. And here we find the categorical dividing line between realisms and anti-realisms or correlationisms: for the anti-realist or correlationist, claims about beings are never claims about beings-in-themselves or beings apart from us, but are always and only claims about beings as they manifest themselves to us. For the realist, by contrast, claims about objects really are claims about objects and not objects as they are for-us or only in relation to us.

As a consequence, it becomes clear that, for the correlationist, objects take on the status of fictions. Because objects can no longer be equated with things-in-themselves, because objects are only ever objects for-us and never things as they are independent of us, objects become phenomena or are reduced to actual or possible manifestations to us. Philosophy now shifts from being a debate about the nature of things-in-themselves or substance, to a debate about the mechanisms by which phenomena are produced or structured. Is it mind that structures phenomena? Language? Power? Intentionality? Embodied experience? Such are but a handful of the options that have been entertained by contemporary theory.

At the heart of correlationism it is thus clear that there is a profound anthropocentrism, for where it is held that being can never be thought in its existence apart from thought, it becomes clear that any claims about being ultimately harbor the implicit colophon that claims about being are claims about being for humans. Moreover, despite declarations of anti-humanism on the part of both Heidegger and the structuralists and post-structuralists, it is clear that these anti-humanisms bring us no closer to realism. For, to put it crudely, what the anti-humanisms object to is not the correlationist thesis of a necessary relation between being and thought such that the two can never be thought apart from one another, but rather the manner in which humanisms situate the primordial correlation in the minds of individual knowers. Thus, for example, structuralist and post-structuralist anti-humanisms emphasize the autonomy and independence of language and social relations. Here the argument runs that it is not sovereign subjects that are calling the shots, but rather language and/or social relations. What the structuralist and post-structuralist anti-humanists wish to examine is the manner in which language and social relations are determinative of the actions of individuals and how, if Althusser is to be followed, the individual itself is an effect of these more primordial agencies. It now follows that these impersonal and anonymous agencies are the condition for manifestation, not individual human minds. World, the story goes, is not a construction of the mind of human individuals or transcendental subjectivity, but of impersonal and anonymous social structures.

However, while anti-humanisms rescue philosophy from its focus on individual minds, allowing us to discern the sway of far more impersonal and anonymous patterns and structures at work in the heart of thought and social relations, it by no means follows that anti-humanism has escaped anthropocentrism. For social relations, economic relations, and language are nonetheless human phenomena, even where the human is discursively constructed, and it therefore follows that we remain within the orbit of anthropocentrism; for just as we began humanistically with the premise that we cannot know what being might be independent of our thinking of being, we now conclude that our claims about being are claims about being in relation to or correlation with language, power, or social relations. The question of what the world might be like apart from humans is, for both the humanists and the anti-humanists, entirely foreclosed.

1.2. Breaking the Correlationist Circle

With correlationism, the question of ontology is no longer, “what is being qua being?” but rather, “what is being qua Dasein?” or, “what is being qua language?” or, “what is being qua power?” or, “what is being qua history?” or, “what is being qua the lived body?” and many other avatars besides. While disputes among these various formulations of the correlation are heated, we are nonetheless faced with a series of anthropomorphic determinations such that being is always to be thought in relation to some aspect of the human. Protagoras, in one form or another, rules the day. What we thus get is not a democracy of objects or actants where all objects are on equal ontological footing and the philosopher can be just as interested in questions of how, to evoke Harman's favorite example, cotton relates to fire as she is in questions of how humans relate to mangoes, but instead a monarchy of the human in relation to all other beings where some instance of the human is treated as that which overdetermines all other beings and where the primary order of the day is always to determine how individual minds relate to other objects or how the social and cultural relates to being. If, as Žižek contends, metaphysics, in the pejorative sense of ontotheology, consists in elevating a part to the ground of the whole, then the anthropocentrism of correlationism is metaphysical through and through despite its protestations to the contrary or its characterizations of itself as a critique of metaphysics.[15] Correlationism is ontotheology with the human in the place of God.

Our question is two-fold: on the one hand, what are the philosophical premises that led correlationism to become such a persuasive position in contemporary philosophy? On the other hand, is there a way to twist free from the correlationist deadlock so as to convincingly defend a genuinely post-humanist, realist ontology? A post-humanist, realist ontology is not an anti-human ontology, but is rather, as we will later see, an ontology where humans are no longer monarchs of being but are instead among beings, entangled in beings, and implicated in other beings. In this section, I will address the second of these questions through a foray into the early thought of the philosopher of science Roy Bhaskar. This discussion will also set the stage for my later discussion of the ontology of objects. Ironically, it turns out that the way out of correlationism is to be found through transcendental argumentation. But here I am getting ahead of myself.

Before outlining Bhaskar's defense of ontological realism, it is first necessary to express a note of caution. In his early work, and especially A Realist Theory of Science on which I will lean heavily here, Bhaskar is primarily interested in the ontology of science. Within the context of the present book, this poses special dangers as readers might arrive at the mistaken impression that I am arguing that the objects of the natural sciences are exhaustive of being as such. In short, one might conclude that I advocate the thesis that being and the objects of the natural world are synonymous. As I develop Bhaskar's argument for a realist theory of science, my aim is not to defend naturalism as the one true ontology, but rather to unfold the argument by which he arrives at a realist ontology. And here I am only interested in the ontological dimension of his argument. As Bhaskar writes,

Bhaskar refers to these two dimensions of knowledge as the transitive and the intransitive respectively. The transitive refers to the dimension of the social in the production of knowledge, such as inherited discourses, scientific training, institutions, and so on. By contrast, the intransitive refers to the domain of being that would exist regardless of whether or not humans know of them. We can thus say that Bhaskar, in his philosophy of science, wishes to reconcile something like the insights of Kuhn's Structure of Scientific Revolutions[17] or Foucault's Order of Things (the transitive or social dimension of knowledge production), with a realist ontology of science (the intransitive dimension). In the present context, my focus is on Bhaskar's arguments on behalf of realist ontology and therefore the intransitive, though readers should not conclude from this that the social dimension through which knowledge is produced is ignored.

As subsequent chapters of this book will hopefully demonstrate, the domain of real being advocated by onticology is far broader than the domain of beings belonging to the natural world. Put differently, natural beings constitute a subset of the category of real beings. In addition to natural beings, onticology also counts technologies, symbolic entities, fictional entities, groups, nations, works of art, possible beings, artificial entities, and many other entities besides as belonging to the domain of real being. Bhaskar's arguments for the reality of natural entities thus functions as a launching point for ontology, but is by no means exhaustive of the domain of real beings.

Consequently, second, the ontology defended in these pages will expand significantly on the ontology Bhaskar proposes in A Realist Theory of Science, going, I believe, well beyond what he proposes in his own ontology. Thus, while I am deeply indebted to the ontology Bhaskar proposes in A Realist Theory of Science, it should not be assumed that the ontology presupposed by onticology is identical to the ontology proposed by Bhaskar's transcendental realism. In what follows, there will be many points of overlap between these ontologies, and many points of divergence as well. My aim in this book is not to provide a commentary on Bhaskar's ontology, nor to remain true to his ontology, but rather to develop a post-humanist, realist theory of being capable of breaking with correlationism, that is nonetheless capable of integrating the most important and significant findings of the correlationists.

Bhaskar's defense of ontological realism begins with a very simple transcendental question: “...what must the world be like for science to be possible?”[18] In asking what the world must be like for science to be possible, Bhaskar is asking a transcendental question and deploying a transcendental mode of argumentation. The question here is not, “how do we have access to the world?” or, “how do we know the world?” but rather what must be presupposed about the nature of the world in order for our scientific practices to be possible. As Deleuze reminds us, the transcendental is not to be confused with the transcendent.[19] The transcendent refers to that which is above or beyond something else. For example, God, if it exists, is perhaps transcendent to the world. The transcendental, by contrast, refers to that which is a condition for some other practice, form of cognition, or activity.

Thus, for example, perhaps speech requires language as its transcendental condition. If this is the case, then it would be because the condition under which it is possible for two people to communicate requires the existence of a shared code in the form of language. The conditions under which it is possible for me to speak to you would here lie in both of us sharing the same language, whether language be something minimal like gestures or something very complex like chains of signifiers capable of self-reflexively commenting on themselves. By contrast, the referent of speech would, perhaps, be transcendent to both language and speech. If the referent of speech is transcendent to speech, then this is because 1) we can speak about fictional entities that have no physical referent or, 2) we can speak of entities in their absence, and 3) we can speak of speech itself, transforming speech into an object. Too often questions of the transcendental have been confused with questions of the transcendent. The point, however, is that transcendental questions are questions about what renders a particular practice or activity possible. Transcendental questions are questions of what a particular practice requires to take place and refer to what is immanent to these practices.

Additionally it should be noted that transcendental questions are not foundationalist in character. Transcendental questions do not seek an absolutely secure and unassailable foundation for knowledge or practice, but merely ask, “given such and such a practice, what must be the case in order for this practice to be possible?” As such, transcendental inquiry sidesteps the epistemological project inaugurated by Descartes and so compellingly critiqued by Hume, by disavowing the project of seeking for an absolute foundation for knowledge.

Yet already, with Bhaskar's question, we sense that the air or atmosphere is very different. Bhaskar does not ask what the mind must be like for science to be possible, but rather what the world must be like for science to be possible. In framing the question of science in this way, Bhaskar shifts the transcendental question from the domain of epistemology to the domain of ontology. The world itself must be a particular way for science to be possible, not the mind. And if Bhaskar's deployment of transcendental argument is ironic as a defense of realism, it is for precisely this reason. For beginning with Kant, who first explicitly invented the transcendental model of argument, all subsequent models of transcendental argument have either traced the conditions of certain forms of practice back to mind or some variant of the social. Yet with Bhaskar we here have a thorough inversion of this mode of argumentation. The question is no longer, “what must the mind be like for X to be possible?” nor, “what must the social be like for X to be possible?” but rather, “what must the world be like for X to be possible?” And in this way Bhaskar already shifts transcendental philosophy out of its constraint to some form of transcendental idealism or anti-realism, and shifts it to the framework of transcendental realism. For in asking what the world must be like for science to be possible, we begin with the premise of a world apart from and independent of human beings.

Thus, elsewhere, in a gloss on the nature of transcendental argumentation, Bhaskar will write,

To this I would add that we must avoid the conclusion that any answer to the question, “what must be the case for ϕ to be possible?” must refer to society, language, or power. In effect, Bhaskar thus proposes a de-suturation of transcendental modes of argumentation from mind and the social. To suture is to tie, and in this respect all correlationism is a suture. The mark of any and every correlationism is a suture of being to the human in some form or another. Thus, a de-suturing would amount to an untying and systematic separation of the domains of being and thought, thereby undermining the reign of Protagoras in the modern era. Yet, nonetheless, as promising as this strategy of de-suturation sounds, we require compelling reasons to follow Bhaskar in this de-suturation of being and thought, or in this move directly from mind to world.

Bhaskar argues that the condition under which science is possible is the existence of what he calls “intransitive objects” which are real structures that exist independently of our minds and that are often “out of phase” with actual patterns of events.[21] As Bhaskar articulates it,

The claim that intransitive objects are invariant to our knowledge of them is not equivalent to the claim that intransitive objects are invariant. Rather, the point is that these objects would do what they do regardless of whether anyone knew about them or perceived them. The claim that intransitive objects can be “out of phase” with actual patterns of events is the claim that these intransitive objects can act or be dormant, thereby not producing certain events that they would produce in other settings or contexts.

So far we know what Bhaskar claims are the transcendental conditions of scientific practice, but we do not know why Bhaskar claims these are the transcendental conditions of science. Indeed, as it stands, it sounds as if he is dogmatically affirming the existence of mind-independent objects without providing any grounds as to why these objects must be presupposed by scientific practice. Here everything turns on Bhaskar's thesis that intransitive objects can be “out of phase” with actual patterns of events and the nature of experiment in scientific practice. “The intelligibility of experimental activity presupposes not just the intransitivity but the structured character of the objects investigated under experimental conditions”.[23] But why is this so? According to Bhaskar,

To understand Bhaskar's point, it is necessary to compare his thesis here to that advocated by sense-data empiricism. “A causal law is analysed in empiricist ontology as a constant conjunction of events perceived (or perceptions)”.[25] Borne out of foundationalist aspirations, the empiricist attempts to eradicate all unfounded ontological presuppositions by resolving to tarry with what is given in sensations alone. For example, for the empiricist there is no deeper fact independent of the sensations. I now regard a clementine with relish. This clementine just is, for the sense-data empiricist, the sheen of orange and roundness I see before me, the fragrant citrus scent that fills my nostrils, the refreshing coolness I feel as I now touch it, and so on. What we call a clementine, for the empiricist, is a bundle of sensations that my mind has unified together through the operations of association. If, according to the empiricist, there is more to the clementine than these sensations afford me, if it has independent existence in its own right, then I can know nothing about it for I only ever have access to the object through these sensations. Therefore the empiricist, practicing a sort of critical vigilance, decides to cut the real independent object out of the discussion altogether as an unjustified metaphysical residue, restricting philosophical discussion to what is given in sensation alone.

When the empiricist arrives at discussions of causality, she thus has no recourse but to discuss claims about causality as claims about constant conjunctions of events or impressions. For the empiricist, a causal claim is no more nor less than a constant conjunction of sensations in experience. Above all, causal claims are not claims about powers that reside in objects, precisely because we have no access to these hidden powers. As Hume puts it, “[t]he bread which I formerly eat, nourished me; that is, a body of such sensible qualities, was, at that time, endowed with such secret powers: But does it follow, that other bread must also nourish me at another time, and that like sensible qualities must always be attended with like secret powers?”[26] What we are given is not the powers of the object—in this case bread—but rather a constant connection of sensations: the visual appearance of the bread followed by the sensations of our painful hunger being satiated. The causal relation is nothing but the mental association of these sensations in the order of time.

We are now in a position to better appreciate Bhaskar's point. Were the empiricist thesis correct, there would be no point to the activity of experiment, for causal knowledge would be nothing but constant conjunctions of events presented in sensations. For the empiricist experiment must be unintelligible precisely because the idea of hidden or disguised powers of objects is banished by the empiricist reduction that resolves to treat being only in terms of what is present or given in sensations. Accordingly, elsewhere, in a moment of humor, Bhaskar will write that

This absurdity would follow from the claim that causal relations are nothing more than constant conjunctions of events or sensation, forbidding any hidden powers in objects, thereby leaving the only site for the emergence of new sensations in the experimenter.

1.3. The Onto-Transcendental Grounds of Experimental Activity

Bhaskar thus concludes that the transcendental conditions under which experimental activity is intelligible are ontological in character. In order for experimental activity to be intelligible, it is the world, not our minds, that must be a certain way. And the world must be this way regardless of whether any science takes place or whether any sentient beings exist to engage in something approximating science. Above all it is necessary that 1) objects be intransitive to our knowledge, perception, or discourse about objects, and that 2) it be possible for objects to be out of phase with actual events. I will address the second of these points first.

If we must draw an ontological distinction between objects and the events they generate to understand the intelligibility of scientific practice, then this is precisely because objects do not ordinarily or regularly produce constant conjunctions of events. Constant conjunctions of events are the exception rather than the rule, and it is for this reason that we engage in experimental practice. In this connection, Bhaskar draws a distinction between open and closed systems. Closed systems are systems where constant conjunctions of events obtain. Open systems are, by contrast, systems where the powers of objects are either not acting or are rather disguised or hidden by virtue of the intervention of other causes. Open systems are the norm rather than the exception. And within open systems or entanglements of objects, the powers of discrete objects are often veiled or inactive. It is here that we encounter the rationale behind experimental activity. As Bhaskar puts it,

We thus have an ontological distinction between objects or generative mechanisms on the one hand, and events, on the other. If experimental activity is necessary, then this is because generative mechanisms can be dormant, inactive, or veiled by the agency of other objects or generative mechanisms. Nonetheless, it is generative mechanisms or objects that are responsible for the production of events. As Bhaskar will remark further on, “[t]he world consists of things, not events. Most things are complex objects, in virtue of which they possess an ensemble of tendencies, liabilities and powers. It is by reference to the exercise of their tendencies, liabilities and powers that the phenomena of the world are explained”.[29]

Because things, objects, or generative mechanisms can be out of phase with events, experimental activity is required to produce closed systems so that the relation between generative mechanisms and events might be discovered. To illustrate this point, Bhaskar provides a diagram distinguishing between the domains of the real, the actual, and the empirical:[30]

The domain of the empirical reduces events to experiences and excludes mechanisms altogether. As we saw in the case of Hume's empiricism, only sensations are given. In the domain of the actual, some events may be experiencable, yet there can be many events that we are yet to experience and even other events that are beyond the possibility of any experience. Moreover, events or actualities can be out of phase with mechanisms, objects, or things. Finally, the domain of mechanisms or the real is inclusive of both mechanisms, events, and experiences without the requirement that these categories overlap or always occur together. Thus there can be the presence of mechanisms without the presence of events or experiences, and there can be the presence of events without anyone about to experience them or anyone even capable of experiencing them. Zebras run across the savannah just fine without the aid of our gaze.

Bhaskar drives his point home rather dramatically by claiming the condition under which experimental practice is intelligible lies in the possibility of “a world without men”.[31] Initially this thesis sounds paradoxical in that it is humans and perhaps other sentient beings that conduct experiments. However, Bhaskar's point lies elsewhere. He is not making the absurd claim that experiment requires no humans or sentient beings to conduct experiments—it does—but rather that because constant conjunctions of events ordinarily require humans or some other sentient being to produce them, because constant conjunctions of events are not the rule but the exception, and because generative mechanisms or objects are ordinarily out of phase with events or actualities, the intelligibility of our experimental activity is premised on the possibility of a world without humans where objects reside in the world unrealized and unwitnessed, and without producing certain actualities such as those we find in the experimental setting.

Consequently, Bhaskar's thesis is radically opposed to something like Lacan's claim that the universe is the flower of rhetoric. Such a thesis is misguided on two grounds: first, this thesis conflates propositions about the world with the world itself. Yet the world requires no propositions about the world to be the world. Second, and more importantly, this thesis renders experimental activity completely incoherent because science does not begin with true propositions, but seeks to create closed systems (where possible) so as to trigger generative mechanisms and thereby produce or uncover constant conjunctions of events. Were the world the totality of true propositions or constructed by language, this activity would be most peculiar indeed as there would be no unknown generative mechanisms to uncover in the experimental setting. In other words, the intelligibility of experimental practice is premised on the ontological supposition of generative mechanisms or objects independent of that activity.

And it is for this reason also that the condition for the intelligibility of experimental activity is the existence of objects that are intransitive or independent of mind and perception. For if objects were dependent on mind, perception, or culture, then there would be nothing to discover in the closed systems produced in the experimental setting. Consequently, not only are intransitive objects the premise of experimental activity, but the generative mechanisms discovered in experimental activity are also treated as operative in open systems once they are discovered, despite the fact that they operate in open systems in a fashion where the events they are capable of producing go unrealized because the mechanism is either dormant or countervailed by other generative mechanisms.

Bhaskar thus argues that claims about generative mechanisms are transfactual. Experimental activity does not show that constant conjunctions of events must always be operative in open systems, but rather that where these generative mechanisms exist without producing actualities or events, they are nonetheless operative in these open systems. Bhaskar thus argues that generative mechanisms must be understood as tendencies or powers. “[T]endencies are potentialities which may be exercised or as it were 'in play' without being realized or manifest in any particular outcome”.[32] Needless to say, a tendency or power is a real feature of objects themselves, a feature of the being of objects, and not the being of objects for-us. Moreover, the distinction between generative mechanisms or objects with their tendencies or powers and events or actualities is an ontological distinction, not a distinction pertaining to our knowledge. Events are real beings or occurrences produced by generative mechanisms, and generative mechanisms are real entities with the power to produce these events. Thus the transcendental conditions Bhaskar uncovers for the intelligibility of experimental practice are thoroughly ontological and realist in character. They are features of the world itself and not the mind that regards the world.

Already we can sense just how far we are from Kant and Hume, both of whom have influenced contemporary theory so deeply, albeit in an often subterranean fashion. Where both Kant and Hume call for an investigation of mind when raising questions of knowledge, Bhaskar calls for a philosophical investigation of the world. For Bhaskar, it is ontology that is first philosophy, not epistemology. More importantly, where both Kant and Hume treat claims about causality as claims about constant conjunctions of events, Bhaskar vigorously rejects the thesis that claims about causality are claims about constant conjunctions of events on the grounds that constant conjunctions of events are the exception rather than the rule, and instead argues that claims about causality are claims about generative mechanisms that may or may not produce certain events depending on their entanglements with other objects.

Additionally, it will be noted that the ontology proposed by Bhaskar is immune to the charge of being a naïve realism. The charge of naïve realism is a favorite lazy rejoinder of anti-realist and correlationist styles of argument whenever encountering a defense of realism. Now, naïve realism is the thesis that the world is exactly as we perceive or experience it. It is, in short, the thesis that the qualities we perceive in an object truly belong to the object itself regardless of whether anyone perceives that object. However, it is clear that nothing could be further from Bhaskar's position. In distinguishing between generative mechanisms or objects and events or actualities, in noting the manner in which objects behave differently in open and closed systems, Bhaskar underlines the manner in which objects are withdrawn from any qualities they might happen to manifest. Put otherwise, a key feature of Bhaskar's argument is that objects or generative mechanisms cannot be equated with or reduced to their qualities. I shall have much more to say about this later when I deal with exo-relations or relations between generative mechanisms or objects, but for the moment it is sufficient to note that the ontology proposed by Bhaskar is anathema to any variant of naïve realism.

1.4. Objections and Replies

No doubt a number of objections will have arisen in the mind of the reader sympathetic to the correlationist line of argument. In particular, three lines of argument can be anticipated: First, that the realist ontology and transcendental line of argument proposed by Bhaskar purports to know objects a priori before knowing them; second, that it is impossible to imagine a world without men because we still imagine ourselves as being present to this world in our absence to this world; and third, in a closely related vein, that it is impossible to think anything without, as it were, including ourselves in the picture of what is to be thought. I will address each of these objections in their turn.

The first of these arguments is the easiest to dispatch. What this argument purports is that realist ontology claims to know before we know. Through this line of attack it hopes to establish the primacy of epistemology over ontology, or that epistemology is, in fact, first philosophy. For, the argument runs, how can ontology make claims about the being of being without first knowing these beings? As Bhaskar formulates this line of argument, “ontology is dependent upon epistemology since what we can know to exist is merely a part of what we can know”.[33] I suspect that this line of argument, more than any other, motivates the subordination of ontology to epistemology and the treatment of ontology as “onto-epistemology”. This seems to follow as obvious: to speak of being, we must know being and therefore an inquiry into knowledge or epistemology must precede any discussion of being. Heidegger, for example, argues that before we can even formulate the question of the meaning of being (it’s noteworthy that he formulates this question as a question of meaning rather than a question of what being is) we must first investigate Dasein's ontic-ontological pre-comprehension of being.[34] Nothing could be more obvious.

However, as Bhaskar argues,

Ontology does not tell us what objects exist, but that objects exist, that they are generative mechanisms, that they cannot be identified with events, actualities, or qualities, and that they behave differently in open and closed systems. These are ontological premises necessary to render our experimental activity intelligible. It is the job of actual inquiry to discover what objects exist. However, if inquiry is to be intelligible then it must begin with the premise that there are objects that act independently of this inquiry.

A second line of argument holds that it is impossible to intelligibly think a world without men because, in the very act of thinking such a world, we are picturing ourselves as present to this world. The thesis here is that every picturing of the world includes ourselves in that picture. However, as Quentin Meillassoux has convincingly argued, such a line of argument leads to the conclusion that the thought of our own death is unintelligible or that we are necessarily immortal. For if it is true that we cannot think the world without thinking our presence to the world, then it follows that even the thought of our own death requires the presence of our thinking, thereby undermining the possibility of dying. As Meillassoux formulates this line of argument, “I can only think of myself as existing, and as existing the way I exist; thus, I cannot but exist, and always exist as I exist now”.[36]

Such is the argument of absolute idealists that 1) denies the existence of an in-itself apart from thought, and 2) argues that the correlation between being and thought is absolute or reality itself, i.e, that there is nothing apart from correlation (Berkeley and Hegel, though in very different ways). In response to this line of argument, Meillassoux cites the agnostic—the correlationist that concedes the possible existence of the in-itself apart from thought while maintaining the index of all thought to phenomena or being-for-us—pointing out that,

While I do not follow Meillassoux in his inference from the contingency of our being to the contingency of being as such (I await a clearer formulation of this argument), I do believe his argument here hits the mark. If it is conceded that our annihilation is, in principle, thinkable, then we are also conceding that a world without humans is thinkable. For to think our annihilation just is to think a world where we are absent. Yet if this is the case, then the correlationist argument that it is impossible to think a form of being apart from thought is severely challenged.

Here it must be emphasized that Meillassoux's argument does not rest on establishing that we are annihilated with death or that we know that death entails the extinction of our being. As unlikely as it is given what we have come to know about the relationship between mind and brain, it could be that we persist after death. All that is required for Meillassoux's argument is that our extinction or annihilation be thinkable as a possibility. And if it is thinkable as a possibility—a point that seems amply supported by people's anxieties about death—then it also follows that it is possible to think a world without humans.

A final, and closely related line of argument, revolves around the reflexivity of thought. Here the idea is that it is impossible to think anything without simultaneously thinking that I am thinking it. Like the second objection, this objection revolves around the thesis that the thinker is always included in the picture of what she thinks. Thus, for example, as I think about making myself a cup of coffee, it is held that I must also be aware of thinking that I am thinking about making myself a cup of coffee. All thought, therefore, must include the thought of the thought in the activity of the thinking. As such, all thought must necessarily be reflexive or must simultaneously be aware of the object that it thinks and the fact that it is thinking this thought. And if this is the case, then it follows that the thinker must be included in the thought of any being other than thought, and that therefore it is impossible to escape the correlation between thought and being.

As Meillassoux amusingly puts it in his Goldsmith's talk, thought turns out to be like a bit of dual adhesive tape one attempts to remove from one’s finger.[38] This talk was the occasion of a conference organized by Alberto Toscano, where Ray Brassier, Iain Hamilton Grant, Graham Harman, and Meillassoux each gave talks. It was here that the term “speculative realism” was used for the first time. In his talk, Meillassoux illustrated the correlationist argument through analogy to a dual adhesive bit of tape stuck to a person’s finger. Each time she attempts to remove the tape it ends up sticking to another finger, such that the tape is inescapable. Likewise, if thought indeed has this character of reflexivity as a characteristic that always accompanies thought, then it follows that correlation or thought is inescapable. Such is the import of Descartes' famous analysis of the wax in the second meditation. As Descartes writes,

Descartes' innocent little thesis here has been the source of much mischief in subsequent philosophy and is one of the root premises of correlationist thought. Whether we are speaking of Kant's transcendental unity of apperception which is treated as something that must accompany all thought[40], or Hegel's dialectical gymnastics where it is shown that the thinker is always included in the thought[41], the root of these claims traces back to Descartes' thesis that all thought is necessarily reflexive. It is this, for example, that will ultimately allow Hegel to assert the identity of substance and subject in The Phenomenology of Spirit.

However, is this seemingly obvious thesis truly so obvious? Meillassoux believes that this is a powerful argument that must be addressed. I’m not so sure. Is it self-evident that any thought must include the thinker or that the thinker is thinking the thought? While I certainly concede the thesis that in many instances we are capable of self-reflexively thinking the thought that we are thinking a thought, I am much more circumspect about the claim that all thought is necessarily reflexive. Were this the case, then it would seem that thought is impossible, for we would fall into an infinite regress. Thus, as I sit here thinking that I would like to make myself a cup of coffee, I would, if the reflexivity thesis were true, have to think that I am thinking that I would like to make myself a cup of coffee. But since it is asserted that all thought is reflexive, I would additionally have to think that I am thinking that I am thinking that I would like to make a cup of coffee, and so on to infinity. Yet if this were what truly takes place in the activity of thought, thought would be paralyzed. As Bhaskar puts it, “[i]t is possible for A to think ε and be aware of thinking ε without thinking about thinking ε; and unless this were so no-one could ever intelligently think”.[42] What we need here is something like Sartre's “pre-reflexive cogito” which thinks something without simultaneously thinking itself.[43] Yet if such a cogito is possible, and indeed it appears necessary, then we have a thinking that doesn't simultaneously posit itself but which is completely absorbed in what we are thinking.

1.5. Origins of Correlationism: Actualism and the Epistemic Fallacy

In 1.2 I raised the question of what philosophical premises render correlationism such an appealing and persuasive hypothesis. We are now in a position to answer this question. In A Realist Theory of Science, Bhaskar identifies “actualism” as the root premise that ultimately leads to correlationism. As articulated by Bhaskar,

Here it should be noted that actualism does not treat the actual as events that take place in the world regardless of whether or not anyone is about to witness them, but rather identifies the actual with what is given in sensations or impressions. Moreover, this hypothesis is not restricted to empiricists such as Hume, but is also carried over by Kant and his descendents. Consequently, it is necessary to distinguish between classical empiricists, such as Hume, and empiricist ontology. Classical empiricists hold that knowledge arises from sensation alone. Empiricist ontology holds that only the actual, construed as what is given in atomistic sensations, is real; or, at any rate, is all that we can speak of.

Kant does not question Hume's thesis that our knowledge is restricted to what is given in impressions or sensations, but rather embraces it wholesale. And because Kant carries over the actualist thesis of empiricist ontology, he is committed to the thesis that causal claims are claims about constant conjunctions of events given in sensation rather than about powers residing in objects or generative mechanisms that may go unactualized. Kant's innovation, therefore, does not reside in rejecting Hume's doctrine of impressions, but in recognizing that psychological operations of the mind such as the principles of association are insufficient to account for the necessity we attribute to causal relations. Sensation, Kant contends, requires supplementation by the mind, for relations are not themselves directly given in impressions. Consequently, Kant will argue that our judgments of necessity arise not from sensations or associations, but rather from the application of a priori categories of the understanding such as cause and effect to the manifold of sensations.

Yet Kant and his heirs are only led to the conclusion that sensation requires supplementation by categories of mind (or culture, language, norms, or power) as a result of presupposing the actualist hypothesis of empiricist ontology. For where knowledge is restricted to the actual, and the actual is equated with sensation or impressions, relations among objects become thoroughly mysterious as these relations are not directly given in the actual. As Harman has persuasively argued, what we get is a secularized form of occasionalism.[45] The occasionalist argues that all events are radically independent of one another. For the occasionalist there is no direct link between objects. In traditional occasionalisms, God is called upon to link objects to one another. Thus, when the paper burns it is not the flame that causes the paper to burn, but rather the intervention of God that brings the paper and the flame into relation with one another.

If, with Hume and Kant, we get a secularized form of occasionalism, then this is because the thesis that events, in the form of sense impressions, are absolutely independent of one another and without relation. As such, mind, rather than God, brings about the relation between events. For Kant, this linkage takes place through the a priori categories of the understanding and the a priori forms of intuition in the form of space, while for Hume this linkage is effected through the operations of association. The problem is that it is unclear how mind acquires this mysterious power to link that which is without linkage and why mind alone should have this privileged capacity. There is no more reason to think that mind should have this power than events themselves. However, this problem only emerges where the real is equated with the actual and the actual is treated as composed of atomistic sensations. Where claims about cause and effect relations are no longer treated as claims about constant conjunctions of events given in experience, but about generative mechanisms that may or may not produce certain events, the problem disappears. Causal claims are not claims about our experience of objects, but about objects themselves. And wherever we encounter arguments to the effect that sensations require supplementation by some other agency such as principles of association, categories of the understanding, language, signs, norms, and so on, we can be sure that actualism is lurking in the shadows and that the ontological conditions for the intelligibility of experimentation have been ignored.

Yet how comes it that we fall into this sort of actualism? Bhaskar contends that actualism originates in what he calls the “epistemic fallacy”. As Bhaskar articulates it,

Earlier Bhaskar remarks that, “[t]hese presumptions can [...] only be explained in terms of the need felt by philosophers for certain foundations of knowledge”.[47]

Here it is necessary to clarify what the epistemic fallacy is and is not about. A critique of the epistemic fallacy and how it operates in philosophy does not amount to the claim that epistemology or questions of the nature of inquiry and knowledge are a fallacy. What the epistemic fallacy identifies is the fallacy of reducing ontological questions to epistemological questions, or conflating questions of how we know with questions of what beings are. In short, the epistemic fallacy occurs wherever being is reduced to our access to being. Thus, for example, wherever beings are reduced to our impressions or sensations of being, wherever being is reduced to our talk about being, wherever being is reduced to discourses about being, wherever being is reduced to signs through which being is manifest, the epistemic fallacy has been committed.

We have seen why this is so, for our experimental practice is only intelligible based on a series of ontological premises and these ontological premises cannot be reduced to our access to being. They are ontological in the robust sense. These ontological premises refer not to what is present or actual to us. Indeed, they refer, as we will see, to beings that are radically withdrawn from any presence or actuality. And as such, they are genuinely ontological premises, not epistemological premises pertaining to what is given.

In recognizing that the epistemic fallacy emerges from foundationalist aspirations on the part of philosophers, Bhaskar hits the mark. It is the desire for a secure and certain foundation for knowledge that leads philosophy to adopt the actualist stance and fall into the epistemic fallacy. These decisions, in turn, ultimately lead to correlationism. In raising the question, “how do we know?” and seeking an argument that would thoroughly defeat the skeptic or sophist, the philosopher concludes that only what is present or given can defend against the incursions of the skeptic. But what is present or given turns out either to be mind or sensations. Therefore the philosopher finds himself in the position of restricting all being to what is given as actual in sensations. From here a whole cascade of problematic consequences follow that increasingly lead to the dissolution of the world as a genuine ontological category.

However, once these foundationalist aspirations are abandoned, the nature of the problem changes significantly and we no longer find ourselves tied to the actualist premise that generates all of these issues. And indeed, these aspirations should be abandoned, for foundationalism is premised on the possibility of absolute presence, absolute proximity, the absence of all absence, and we have now discovered that it is being itself that is split between generative mechanisms or objects and the actual. Difference, deferral, absence, and so on are not idiosyncracies of our being preventing us from ever reaching being, but are, rather, ontological characteristics of being as such. Moreover, this split at the heart of all beings is not simply characteristic of those objects that we would seek to know, but are also characteristics of the peculiar object that we are. We ourselves are split. If, then, this split is a general ontological feature of the world, then the dream of presence required for any form of foundationalism is a priori impossible. We are then left with two paths: to persist in the correlationist thesis that would reduce ontological questions to epistemological questions and which is itself implicitly premised on the ontotheological assumption of actualism, or to investigate the split in being in a post-humanist, realist fashion that is genuinely ontological. It is the second of these two paths that I here attempt.

1.6. On the Alleged Primacy of Perception

In response to the preceding line of argument, one might argue that nonetheless we must identify objects in order to know the being of objects, and that the identification of an object requires some reference to a perceptual convention. To be an object is to possess a boundary or to be distinguished from other things, and in order to distinguish one object from another we must refer to perception. This, then, would be the first step in the argument, asserting the primacy of epistemology over ontology. The next step would then consist in pointing out that different creatures perceive or divide up the world in different ways. Thus, for example, while I very clearly see a tree, it is unlikely that the amoeba encounters this tree as a tree. Likewise, flowing water encounters no difference between a frog jumping into its currents, a rock obstructing its path, or a gust of wind impacting it. The point, then, is not only that we must refer back to epistemology or perceptual conventions to distinguish objects, but that also, while there is something other than perception (Berkeleyian subjective idealism is mistaken), there is no reason to suggest that being-in-itself is composed of objects or that this something is anything like our perception of the world.

This line of argument underlines just why it is so important to distinguish between arguments advanced on ontological grounds and arguments advanced on epistemological grounds. Why is this distinction so important? The reigning assumption in philosophy since the 17^th century is that questions of epistemology must precede questions of ontology. The idea here is that we must first know an object before we can begin talking about the being of objects. This hearkens back to common readings of Meno's paradox in Plato. In the Meno, Socrates asks, “how can we inquire into the nature of virtue without first knowing virtue?” And if this constitutes a paradox, then this is precisely because if we already know virtue, then we have no reason to inquire into the nature of virtue.

The question of perception is not a question about the being of objects, but a question about our access to the being of objects. The point of the question is two-fold: first, the claim is that in order to talk about the being of objects we must first have access to objects. Second, the claim is that perhaps our access to objects has nothing to do with what reality itself is like. This is the point of the amoeba and the tree. The amoeba doesn't encounter the tree as a tree, and thus we should be skeptical of the idea that entities like trees are independent or real entities at all. The thesis is thus that the being of an object arises not from the object's own independent structure, but rather from the distinctions the being perceiving it makes. This is the correlationist gesture par excellence. To be sure, the correlationist may concede that there is something other than the amoeba, but he wishes to argue that there's no reason to suppose that this something is anything like how the amoeba experiences it because the nature of the being that the amoeba perceives is a function of the amoeba's distinctions, not of the being of this other-being itself.

Let us now return to the difference between arguments that are ontologically driven and arguments that are epistemologically driven. The first point to note with respect to the correlationist's argument is that it seems to ignore the fact that this argument already concedes the existence of at least one object. What object is that? Certainly not the existence of the tree. Rather, the correlationist concedes the existence of the amoeba. In order for the amoeba to grasp anything as anything at all, it must exist as an entity, substance, or object. In short, the correlationist's argument can only get off the ground through the presupposition of at least one entity. And this is a central reason that arguments about how observers constitute objects are unconvincing: these arguments always forget that the observer is an object.

Let us return to the example of the water encountering a rock, frog, or gust of wind. Here the argument was made that these entities aren't distinct for the water but produce all the same qualitative effects. In chapter four we will see why this is the case, but for the moment, rather than speaking about the water, let's instead speak of the relation between a frog and the amoeba. When we discuss the relation between the amoeba and the frog, we encounter exactly the same problem as the relation between the water and stones, frogs, and wind: the frog does not distinguish amoebas. The amoeba does not encounter the frog as a frog, nor does the frog encounter the amoeba as an amoeba. From the standpoint of the frog's experience, the amoeba is indistinguishable from air or water. The frog is every bit as indifferent to the existence of the amoeba as an amoeba, as the water is indifferent to the existence of a rock or frog or the wind as a rock, frog, or wind. There is no real difference here. It might as well not exist. However, here's the rub: does the fact of the amoeba's non-existence within the frog’s Umwelt have anything to do with the amoeba's existence? To claim that it does is to be led to the peculiar conclusion that it is the entity's distinctions that make other entities what they are.

Here, then, we arrive at the difference between epistemologically-driven arguments and ontologically-driven arguments. Epistemologically-driven arguments will always pitch questions of what beings are in terms of our access to these entities. Rather than asking what beings must be by right, we instead ask what conventions we use to distinguish entities. However, as can be clearly seen, this changes the issue or question. Rather than treating the question as a question of what beings are, the correlationist instead transforms the question into a question of how we know what things are. And because the correlationist has transformed the question from the issue of what must belong to entities by right in order to be entities, regardless of whether anyone else knows these entities, to the question of how we know entities, the correlationist finds himself confronted with the question of givenness or access. Having transposed a properly ontological question into an epistemological question, and having thereby arrived at the problem of givenness (the reference to perception) or access, the correlationist now notices that different entities or observers perceive the world differently, i.e., that the world is given in different ways to different observers. The amoeba doesn't encounter the tree as a tree. A person who is colorblind cannot see the color purple, and so on.

Based on this line of argumentation, we can now see why Bhaskar refers to the epistemic fallacy as a fallacy. The epistemic fallacy does not lie in raising questions of epistemology. That would be absurd. Of course we should raise epistemic questions. The epistemic fallacy consists in the thesis that properly ontological questions can be fully transposed into epistemological questions. Because the correlationist has transformed questions of what beings are into questions of our access to beings, and because questions of access necessarily trace back to questions of givenness, givenness now comes to legislate what exists and what does not exist. The correlationist is therefore compelled to argue that the tree does not exist or that frogs and rocks don't exist. If this move is problematic, then this is because it always finds itself trapped in a self-referential paradox: to wit, it concedes the existence of at least one entity, and then uses the manner in which that entity observes the rest of the world through its own distinctions to erase the existence of other entities. Every argument of this sort, driven by how we cognize or perceive the world, will run afoul of this sort of problem.

The point here is that questions of ontology cannot, in any manner, shape, or form, be reduced to questions of epistemology. Put otherwise, claims about the being of beings cannot be transposed into claims about our access to beings. Wherever claims about the being of beings are transposed into questions about our access to beings, we end up with givenness legislating what exists and what doesn't exist based on what is given or accessible, and we thereby find ourselves trapped in a self-defeating self-referential paradox where we simultaneously concede the existence of objects while denying their existence.

As a consequence, claims about the being of entities are arrived at in an entirely different manner than the epistemological question of access. We do not begin with our access to beings, but instead ask what the world must be like for certain practices to be possible. The object-oriented ontologist is not claiming that we have access to beings, that they are given, or that our perception is identical to the way the world is, but that the existence of substance is a necessary premise for a whole slew of our practices to be intelligible. In other words, the onticological thesis is that the world must be a particular way for certain practices and activities like perception, experimentation, discourse, and so on, to be possible and that the world would be this way regardless of whether we perceived, experimented, or discoursed about it.

Perhaps the best way to defeat the correlationist is to shift the terms of the debate. It is almost always the case that the correlationist proceeds through a discussion of how humans perceive and discourse about the world. Rather than beginning with humans, however, why not instead begin with the amoeba? Does the correlationist really wish to claim that the amoeba constitutes his being? This conclusion follows directly from the correlationist argument about how the amoeba encounters the tree. If he doesn't wish to arrive at this conclusion, then why? There are only two possible conclusions here, both of which lead to the collapse of the correlationist's argument. The first possible conclusion would be that it is not possible for the amoeba to constitute the correlationist's being because humans are somehow special in the order of being by virtue of being the only beings capable of constituting other beings from the primordial flux of “other-being”. The second possible argument is that the amoeba doesn't constitute the correlationist's being through perceiving the correlationist because the correlationist is a substance or independent being in his own right and how something perceives another being has nothing to do with that being's status as a substance.

Now, one might expect the realist to reject the first possibility on the grounds that it is anthropocentric. However, while this is true, this is not my argument. If we follow the correlationist in the first counter-argument (which really is the disavowed, yet fully embraced, implicit premise of all correlationisms), we have to note that the correlationist has conceded the existence of at least one object, namely, the correlationist himself. From here it is but a short step to asking why humans or the correlationist should have this privileged status within the order of being. Moreover, it's quite remarkable that any being should be able to perform this feat like Atlas holding the world on his shoulders, carving up a structureless world, a pre-individual flux, into discrete packets or units. If we grant the second argument, then, of course, we've conceded the existence of withdrawn substances that have their own being regardless of how other substances perceive them. Those who advise us to observe the observer somehow seem to miss the point that the very act of observing the observer or observing how other observers observe presupposes the existence of an observer that is doing the observing of other observers. Far from undermining the thesis that substances or objects exist, in other words, this move presupposes the existence of at least one substance or object. And as a consequence, this move is incapable of consistently maintaining the thesis that the world is a product of how observers perceive other objects.