Correlationism – An Extract from The Meillassoux Dictionary


Written by Levi R. Bryant

Meillassoux’s concept of correlation is arguably among his most significant and controversial contributions to philosophy. In After Finitude, he defines correlation as ‘the idea according to which we only ever have access to the correlation between thinking and being, and never to either term considered apart from the other’ (AF, 5). Although Meillassoux does not himself specify this, correlationism presumably comes in a variety of different forms, and is therefore not restricted to theories focused on the relation between mind and being. Thus the relation between transcendental ego or lived body and the world in phenomenology would be one variant of correlationism, while the relation between language and being in Wittgenstein, Derrida and Lacan, or between power and knowledge in Foucault, would be other variants. In each case we encounter the claim that being cannot be thought apart from a subject, language or power.

Meillassoux argues that correlationism has been the central notion of philosophy ever since Immanuel Kant, whose core epistemological hypothesis is twofold. On the one hand, Kant argues that objects conform to mind, rather than mind to objects. Kant claimed that in traditional forms of epistemology the mind was conceived as a mirror that reflects being as it is in-itself, independent of us. He argues that mind does not merely reflect reality, but rather actively structures reality. Consequently, on the other hand, he argues that we can never know reality as it is in itself apart from us, but only as it appears to us. If the mind takes an active role in structuring reality (for us) we are unable to know what it is in-itself because we cannot determine what, in appearances, is a product of our own minds and what is a feature of things as they are in themselves. This is because we cannot adopt a third-person perspective that would allow us to compare things as they appear to us and things as they are in themselves. Consequently, knowledge is restricted to appearances and we must remain agnostic as to what being might be like in itself.

The claim that modern philosophy is inspired by Kantian correlationism is not the claim that most modern philosophers embrace the specific details of Kant’s philosophy. Clearly Wittgenstein, for example, does not adopt Kant’s account of transcendental categories, pure a priori intuitions, or the transcendental ego when he speaks of language games. Rather, the correlationist gesture consists solely in the claim that we can only think the relation between being and thinking and that therefore our knowledge is restricted to appearances.

Correlationism is not merely the thesis that we must relate to something in order to know it. Obviously we had to discover dinosaur fossils to know anything of the past existence of dinosaurs. For Meillassoux, it is not relation per se that makes a position correlationist, but rather the assertion of a very specific, unsurpassable relation. As Meillassoux remarks:

Correlationism rests on an argument as simple as it is powerful, and which can be formulated in the following way: No X without givenness of X, and no theory about X without a positing of X. If you speak about something, you speak about something that is given to you, and posited by you. Consequently, the sentence: ‘X is’, means: ‘X is the correlate of thinking’ in a Cartesian sense. That is: X is the correlate of an affection, or a perception, or a conception, or of any subjective act. To be is to be a correlate, a term of a correlation . . . That is why it is impossible to conceive an absolute X, i.e., an X which would be essentially separate from a subject. We can’t know what the reality of the object in itself is because we can’t distinguish between properties which are supposed to belong to the object and properties belonging to the subjective access to the object. (SR, 409, my emphasis)

Correlationism is thus not the thesis that we must relate to something in order to know it, but rather that what we know of anything is true only for us. In this regard, correlationism is a form of scepticism for it asserts that whether or not things-in-themselves are this way is something we can never know because we can only ever know things as they appear to us, not as they are in themselves. For example, for the correlationist there is no answer to the question of whether carbon atoms exist apart from us and whether they decay at such and such a rate because we only ever know appearances. This is Meillassoux’s support for scientific realism. For the correlationist we are never able to get out of the correlation between thought and being to determine whether or not carbon itself has these properties or whether it is thought that bestows these properties, which is sometimes the view of scientific functionalism. Meillassoux calls this unsurpassable relation the correlationist circle.

One of Meillassoux’s central projects lies in finding a way to break out of the correlationist circle. He seeks to determine whether it is possible to think the absolute or being as it is in-itself apart from mind, and what characteristics the absolute might possess. Meillassoux’s discussion of ancestrality or statements about time prior to the existence of human beings is not an argument against correlationism per se, but is designed to present readily familiar and widely accepted claims about cosmic time prior to the existence of life and humans that ought not be permissible within a correlationist framework. If correlationism is true, what entitles us to make claims about the nature of the universe billions of years prior to the emergence of life or mind? Meillassoux presents his account of how we might break out of the correlationist circle in his discussion of the principle of factiality in After Finitude.

The Meillassoux Dictionary Correlationism by Levi Bryant is an extract from The Meillassoux Dictionary edited by Peter Gratton and Paul J. Ennis, published by Edinburgh University Press.

Levi Bryant is Professor of Philosophy at Collin College outside of Dallas, Texas. He is the author of Onto-Cartography: An Ontology of Machines and Media in the Speculative Realism series published by Edinburgh University Press.

Articles: 162


  1. Levi,

    I wonder if Meillassoux is right about Kant. I’m really trying to understand both better.

    You say that Kant says our mind actively structures reality rather than passively mirroring it back to the mind. But already we may be using the word “reality” equivocaly. In the first sense—“mind actively structures reality”—Kant is concerned with epistemology, so reality here refers to the representation of reality in the mind. This is structured by concepts and intuitions. In the second sense we are talking about “reality” as it really is. But unless I’m really wrong about Kant—which is possible—Kant never says our minds are actually structuring objectively real reality, the world of objects in themselves.

    But we can generalize this concept to apply to the hammer and the nail. When the hammer hits the nail it can ever only know the nail as a hammer. It cannot know the nail as a board, or as a carpenter, or as air, or as a bird. And when it strikes the nail that striking does not exhaust all that it is to be the nail. The nail, as an object, is so very very real that it exceeds all its relations and interactions. Likewise, when a human being knows a nail they can only know the nail as a human. We can never know the “thing in itself” because 1) the object has a depth of meaning that isn’t fully exhausted by our knowing it, objects recede from each other (Graham Harman), and 2) we are knowing it in a particular way, as human beings with human minds that have certain cognitive habits, limitations, and categories. I don’t know what arguments there are for denying these two points.
    So correlationism is the error that says, “we can only ever know the correlation of world and mind,” just as the hammer only ever knows the correlation of nail and hammer and nothing more. What exactly is wrong with that? What is the alternative? That our knowing fully penetrates to the bedrock reality of everything we turn our attention to? Or, more modestly, we have some kind of access to just a bit of real reality in an extra ordinary sense? Please help me understand.

  2. Levi, Any insight you can share concerning my questions about correlation would be appreciated. Is the problem with correlationism that Meillassoux thinks we can actual know something beyond the conditions of our human ability to know and our particular human mode of knowing? What am I missing?

  3. Nope, Meillassoux is about Kant is not, apparently, even remotely serious. I will add that there is even a degree of imprecision in Meillassoux’s treatment of Hume. Note that to read Meillassoux at all, you have to find relevant his problematic of “ancestral” times and the question of the “arche-fossil,” the point at issue being what sense we should make of scientific statements about events anterior to the emergence of life on Earth. It’s a point Meillassoux considers, it’s, like, huge for him — Meillassoux’s discussion of ancestrality or statements about time prior to the existence of human beings. He gets really hot about how true, are readily familiar and widely accepted claims about cosmic time prior to the existence of life and humans. Well, do you find this relevant?

Leave a Reply

Your email address will not be published. Required fields are marked *