Internal Realism and Correlationism

l'eternite_courbetOver at Agent Swarm, Terrence Blake claims that Quentin Meillassoux’s notion of correlationism  is excessively narrow since it disqualifies realist positions which respond to worries about access, objectivity and truth raised by transcendental philosophers from Kant through to Husserl, and Heidegger. I’m not sure if Meillassoux’s speculative solution works and I share his worries about Harman’s OOO. But I don’t see any reason to doubt that  the concept “correlationism” beautifully describes a range of contemporary anti-realist philosophies, not all of which are written in the  house style of the post-Kantian European tradition ((Kant, Hegel, etc.). Hilary Putnam’s internal realism is a particularly salient example of correlationism within the pragmatist/analytic camp because it wears its Kantian heart on its sleeve.

Internal Realism is a philosophical oxymoron since it denies that there are things whose existence and nature is independent of human descriptive practices. The fact that Putnam expresses his variant of transcendental philosophy in the post-Wittgensteinian argot of linguistic practices and language-games rather than transcendental subjects or Daseins is largely irrelevant since the roles that language and subjectivity play in correlationist philosophies are, to put it bluntly, correlative (Perhaps, as Frank Farrell argues, “language” and subjectivity” are a hangover from the Nominalist God whose omnipotence extended to determining differences and similarities within an unstructured universe – See Farrell 1996). Meillassoux does not address analytic correlationism in After Finitude but his formulation of correlationism seems to apply to post-Wittgensteinian position for which language and practice assumes the mantle of the transcendental subject:

In the Kantian framework, a statement’s conformity to the object can no longer be defined in terms of a representation’s ‘adequation’ or ‘resemblance’ to an object supposedly subsisting ‘in itself, since this ‘in itself is inaccessible. The difference between an objective representation (such as ‘the sun heats the stone’) and a ‘merely subjective’ representation (such as ‘the room seems warm to me’) is therefore a function of the difference between two types of subjective representation: those that can be universalized, and are thus by right capable of being experienced by everyone, and hence ‘scientific’, and those that cannot be universalized, and hence cannot belong to scientific discourse. From this point on, intersubjectivity, the consensus of a community, supplants the adequation between the representations of a solitary subject and the thing itself as the veritable criterion of objectivity, and of scientific objectivity more particularly. Scientific truth is no longer what conforms to an in itself supposedly indifferent to the way in which it is given to the subject, but rather what is susceptible of being given as shared by a scientific community.

Such considerations reveal the extent to which the central notion of modern philosophy since Kant seems to be that of correlation. By ‘correlation’ we mean 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. We will henceforth call correlationism any current of thought which maintains the unsurpassable character of the correlation so defined (Meillassoux 2006, 4-5).

Putnam is a modern Kantian because he regards ontology as internal to languages or conceptual schemes (though, for Putnam, unlike Kant, these categorical frameworks are historically contingent). There are no ontological facts that obtain independently of some fixation of language. Such facts would require the existence of a One True Theory of reality which, he claims, is precluded on model theoretic grounds:

The suggestion I am making , in short, is that a statement is true of a situation just in case it would be correct to use the words of which the statement consists in that way in describing the situation. Provided the concepts in question are not themselves ones which we ought to reject for one reason or another, we can explain what ” correct to use the words of which the statement consists in that way ” means by saying that it means nothing more nor less than that a sufficiently well placed speaker who used the words in that way would be fully warranted in counting the statement as true of that situation (Putnam 1987, 115).

As a number of commentators have argued the semantic considerations that motivate Putnam’s shift from realism to internal realism are precisely the one’s that motivated Kant to develop a non-representational account of concepts (See Moran 2000).  While Putnam is exemplary, similar considerations apply to Dummett-style anti-realism. Davidson is a harder case because, unlike Putnam, Davidson rejects epistemic accounts of truth (Davidson 1990, 307-9). However, Davidson thinks that what Tarski leaves out when he shows us how to determine the extension of the truth predicate relative to an object language L is a presupposition of our intersubjective practices of interpretation. Thus, as Jeff Malpas argues, Davidson is probably some kind of “horizontal realist” for whom the world must be understood as the open phenomenological background against which interpretative practices operate – thus looping us back to transcendental subjectivity in its most developed, subtle but still humanist formulation. Horizontal realism is still realism with something missing. It is not relativism, strictly speaking, but the “world” that it presupposes is more like Husserl’s pre-theoretically given Lebenswelt than Meillassoux’s great outdoors (Malpas 1991)



Davidson, Donald (1990). The structure and content of truth. Journal of Philosophy 87 (6):279-328.

Farrell, Frank (1996). Subjectivity, Realism and Postmodernism: The Recovery of the  World in Recent Philosophy ( Cambridge University Press).

Malpas, J.E. (1992) Donald Davidson and the Mirror of Meaning. Cambridge: Cambridge      University Press.

Moran, Dermot (2000). “Hilary Putnam and Immanuel Kant: Two `internal realists’?” Synthese 123 (1):65-104.

Meillassoux, Q. (2006) After Finitude: An Essay on the Necessity of Contingency, Ray Brassier (trans.). New York: Continuum.

Putnam, Hilary (1987). Representation and Reality. MIT Press.

8 thoughts on “Internal Realism and Correlationism

  1. I think that terms such as “realism”, “anti-realism” etc. are very slippery and often change meaning several times in the course of a discussion. For example, Harman makes a big fuss about his “realism”, yet he is anti-realist with respect to scientific objects and common sense objects. His vanishingly thin realism is what I call tautological realism, the only real entities are his real objects (or “harms” as you once called them) but he can give no example as they are unknowable, untouchable, invisible etc.
    I do not find Putnam very interesting as he came rather slowly to his own rather constipated position of “internal realism” but I think that he just does not fall under Meillassoux’s concepts. What I find ridiculous in Meillassoux’s definition is the confusion of knowledge and access. Knowledge is not access, more generally “aboutness” is not access or attempted access either. Similarly “thinking” is not the same thing as “theory”. Meillassoux’s little anti-correlationist machine cannot even get started on most Anglophone epistemologies.
    The term “correlation” has a sort of intuitive appeal, but that is precisely its danger. I have nothing against intuition as such, but it is a simplistic sort of foundation for a critique of all major post-Kantian philosophies. Meillassoux seems to be unaffected by all the debates on theory-change and scientific revolutions, the theory-ladenness of observation, the existence of normal science, the historicity of mathematics, the role of mediations etc.
    Putnam himself was very slow to come to his position and always seemed to me to be conceptually very conservative. Yet he is miles ahead of Meillassoux. I have, however, no particular wish to defend Putnam, and I find, as you do, that his notion of the scientific community is ambiguous, coming very close to being just a bigger subject, a group subject. It leaves out the necessary mediations. Or does it? It all depends on what he has packed into the notion of “warrant”. Is it just a linguistic notion? I think not. Observations, tests, laboratories, crucial experiments all play an essential role. Internal realism is not “debating club” realism. Nor is it the thesis of the infinite plasticity of the real to our theories. That is an empirical thesis, overwhelmingly refuted by the facts under any warranted description.

  2. Hi Terrence, I accept that “realism” and “anti-realism” are loaded terms – every philosophical term is. Putnam’s metaphysical realism is a kind of representational realism that many soi-disant realists like Delanda or John Heil would not accept.

    But I’m less interested in what Meillassoux leaves out (point taken, again) than in what he leaves in: and that’s a diagnostic tool which captures something phenomenological positions like Husserl’s and certain post-Wittgensteinian positions patently share. Nor am I defending Meillassoux’s speculative response to correlationism, as should be obvious.

  3. One of the biggest weaknesses (for me at least) is that the Speculative Realist group & related offspring say so little about truth; better qualify that with a ‘as far as I know’.
    I find it especially weird as one of the first steps to escape correlationism is to reject the idea of truth as an epistemic category. If you want to claim to be a realist but don’t think truth is a relation between language-or-subject and world, then you’re a realist Jim, but not as we know it …

  4. I suppose there are a number of issues here, Jon. For starters, whether the concept of truth is or ought to be epistemic (defined in terms of some ideal of confirmation or justification); secondly, whether realism consists in a semantic thesis about truth. Putnam thinks that the metaphysical realist is committed to a semantic thesis about truth. This is where the model theoretic argument bites unless you can motivate the claim that there are objective semantic relations out there (e.g. some kind of causal relationship between words and things) which determine which model for a theory is the intended one independently of its logical articulation.

  5. David, I feel that we are talking at cross purposes and that we may agree more than you think. On Putnam I agree that to the extent to which his realism is representational, and not constructivist, he is to be criticised. I am talking about the “internal” side of his realism. I do not think however that Putnam believes that with the passage from geocentrism to heliocentrism the orbits of the planets actually changed, or that what we meant by “planet” and “orbit” became so different that the two theories could not be compared. I think that causal chains of reference trump internal chains of interpretation. One consequence would be that “warranted” is time-dependent and so susceptible to degrees. A modern post-Galilean heliocentrist is more warranted now in his astronomical assertions not only than a modern geocentrist but also than ancient pre-Copernican geocentrists. Some of what Galileo said, although it is accepted and demonstrated now, was not warranted at the time he said it, because he did not have adequate auxiliary theories, for example a good theory of optics to legitimate his telescopic observations. This is where I disagree with epistemologies of warranted assertibility, as warrant itself is constructed sometimes by means of unwarranted leaps.
    As to Meillassoux we will have to agree to disagree. I see nothing new in his concept of “correlationism”. But this in itself is not a problem, except for thev amazing propaganda-machine that would have us believe in the impôrtance of Meillassoux’s thought. The real objection is that it represents a tremendous regression behind the last 60 years, at least, of Anglophone and Continental philosophy. As soon as one recognises the constitutive role of objectivated theory in the structure of knowledge (Popper’s Third World inscribed in books and computers and the strucure of instruments and laboratories, Stiegler’s tertiary retentions beginning with writing through print and analogical media up to digital technologies) the correlationist circle cannot be formed, we are Outside and always were. This notion of correlation is for me a gross travesty of the history of thought. You are right for Husserl, but Stiegler argues convincingly that with his late ORIGIN OF GEOMETRY he effectuated already, before Meillassoux was even born. So as a diagnostic tool it is far more limited than OOOxian sweeping claims would have us believe. And I am sorry to insist that the history of Anglophone epistemology has nothing to do with questions of “access”.

    1. As you say, the disagreement is mainly about the diagnostic utility of the correlationism idea, so I won’t say anything further. Steigler on Husserl sounds interesting – have you a reference ?.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s