Fictionalism about Science and the Problem of Ancestrality

Version of June 21, 2023

This English version is a machine translation of the Japanese original, made with Claude (Anthropic). Quotations were translated from the Japanese editions cited, so they are translations of translations and do not match the original texts or their published English translations.

In his book After Finitude [1], Quentin Meillassoux claims that “correlationism” pervades modern and contemporary philosophy, and as a critique of it he raises “the problem of ancestrality” (Chapter 1). Meillassoux holds that correlationism interprets science fictionalistically (ibid., p. 29). In this essay, I take Meillassoux’s argument as pointing out a difficulty that arises when “ancestral statements” are interpreted fictionalistically. And I hold that this difficulty lies not in the temporal “distance” of ancestral statements (ibid., p. 37), but in the fact that science envisages the possibility that the “donation” of the given (ibid., p. 40) might not have been realized.

Correlationism as Meillassoux Understands It

Meillassoux criticizes correlationism, which he takes to pervade contemporary philosophy widely, and tries to demonstrate “speculatively” that there are facts that hold independently of the existence of human beings or other epistemic subjects. Correlationism is the position which

not only says that we can never grasp the object “in itself”, separated from its relation to the subject, but also maintains that the subject is always already placed in a relation to the object, and that we can never grasp a subject that is not. ibid., p. 16

and it is said to encompass Kant’s critical philosophy, transcendental phenomenology, the theory of the life-world, and so on. Whether these philosophical positions really fit what Meillassoux calls correlationism is debatable, but this essay does not take up that question. Still, if we formulate correlationism as “the holding of certain kinds of facts always presupposes the existence of an epistemic subject, and such facts are, in some sense, facts in the intrinsic sense,” I think this is not far off the mark, both as the kind of summary these philosophical positions tend to receive when they circulate in popular form, and as an idea that lies in their background even if it may not be strictly asserted.

The Problem of Ancestrality

Now, Meillassoux criticizes correlationism by means of the concept of “ancestrality.” A fact is said to be ancestral when it is “prior to the emergence of the human species − and also prior to every form of life on earth, as far as can be known − “ (ibid., p. 24). Contemporary science makes a great many statements about ancestral facts, for example “the Earth was born 4.6 billion years ago,” and the criticism is that correlationism cannot legitimately interpret such statements. For the correlationist “says that we can never grasp the object in itself, separated from its relation to the subject” (the passage quoted above). If that is right, objects that precede the subject should be impossible to grasp, and so it should be impossible to grasp ancestral facts as well.

The Correlationist’s Fictionalist Interpretation of Science

However, as Meillassoux points out, most correlationists accept the legitimacy of scientific statements. The strategy correlationists adopt for this purpose is to regard science as a kind of fiction. That is, the correlationist holds that scientific statements do not literally mean what they say, but are statements within a certain fiction, and do not represent any fact. Nevertheless, the correlationist holds, for example, that the use of scientific statements is justified because they are useful.

“Event X occurred many years before the emergence of humans.” (…) The correlationist philosopher will quietly add to the end of the sentence − probably doing so in his own mind − something like a simple codicil-style proviso, always the same one: namely, event X occurred many years before the emergence of humans − for humans (or, to go further, for scientists). ibid., p. 29 In other words, the correlationist accepts scientific statements by paraphrasing them as “Within the fiction called science, ‘the Earth was born 4.6 billion years ago.’”

Meillassoux, however, criticizes this. One of his reasons is what analytic philosophy calls the attitude problem ([2], p. 121).

If the value of an ancestral statement could be drawn only from the fact that its verification is universally possible in the present, then the statement would be of no interest to the scientists who labor to establish it (…) Science does not aim, through its various experiments, to establish the universality of experiment. Science aims, through reproducible experiments, at an external referent that gives meaning to the experiments. [1], p. 35 Scientists are not investigating fictions but facts; otherwise the significance of science could not be explained.

In this essay, however, I take up not this argument but Meillassoux’s rejoinder (ibid., p. 39) to the correlationist’s objection to Meillassoux, and I would like to put it into a more formal shape.

The Necessity and Contingency of the Existence of Epistemic Subjects

First, let me make the claim of correlationism precise. Although Meillassoux does not make this explicit in the text, the correlationist can be taken to assume a collection of statements about objects that the subject can grasp, that is, a collection of statements that are in some way directly and intrinsically meaningful. For such an intrinsically meaningful statement P, correlationism can be taken to assert

If P, then there exists an epistemic subject I, and I cognizes P (the postulate of correlationism)

In the quotation from Meillassoux, ibid., p. 16, only the grasping of P may seem to be at issue, but that would make the claim trivial, so I summarize it in this way.

By contrast, for the correlationist, many scientific statements that cannot be directly experienced, such as those about phenomena occurring in the Andromeda galaxy, are not included in the “set of directly and intrinsically meaningful statements”; they are merely “schematic givens” (ibid., p. 37) that are in some sense constructed from, or verified by, such statements. In this essay, I take it that certain kinds of scientific statements are translated into directly and intrinsically meaningful statements. For although science may not be directly and intrinsically meaningful, it can be thought ultimately to make some statements about givens that are, in a certain sense, “directly given” to us. Since the correlationist accepts the claims of science, the correlationist accepts

If science asserts S, which is translated into an intrinsically meaningful statement P, then P (the transfer principle)

Now consider an ancestral statement, for example “life arose from a chance combination of RNA.” According to Meillassoux, the point of this statement is neither its “distance” from the given, nor the fact that no subject existed at the time to observe this fact; rather, it is that the statement raises the problem of “the time in which the given itself passes from non-being into being” (ibid., p. 41). Still, this statement of Meillassoux’s is not easy to understand. For we already live at a point in time at which the given exists. And the correlationist holds that, before the given existed, there were, strictly speaking, no facts. There seems to be no contradiction here.

I would like to think of it as follows. If there was a time when the given itself did not exist, then it is possible that the given might never have been given at all. For example, had the rate of expansion of spacetime at the Big Bang been slightly different, epistemic subjects could not have existed in this universe, or something of that sort. I think that the problem of ancestrality arises not because it concerns the past, but because it asserts that epistemic subjects have a contingent origin, and hence that their existence is contingent.

I do not claim in this essay that this interpretation is strictly Meillassoux’s own position, but some evidence for its legitimacy can be pointed out. Meillassoux speaks of “Time with a capital T” (ibid., p. 107). This is something that has the potential to destroy even all the laws of nature, but it seems to me that Meillassoux thinks this time ultimately coincides with physical time. For since Meillassoux is trying to defend a realist interpretation of the natural sciences, the time that physics deals with must be the time that really exists. Hence “the time in which the given itself passes from non-being into being” is the time dealt with by the natural sciences, and the “donation” of the given (ibid., p. 40) is an event within the time dealt with by the natural sciences. Now consider an event, for example the event of conscious beings arising in the universe. Unless that event comes about in a completely deterministic way, an event’s having its origin in time means that the event might not have occurred. The “time” Meillassoux has in mind here is in the first instance “Time with a capital T,” but it is real time, and in particular it must be what physics describes. Hence the time before conscious beings, which the natural sciences infer from the “arche-fossil” (ibid., p. 24), entails that conscious beings are contingent.

Now, what consequences does it have that science asserts that the existence of epistemic subjects is contingent? Since the existence of an epistemic subject seems to be a directly graspable fact, the transfer principle seems to show that the existence of epistemic subjects is contingent. On the other hand, the postulate of correlationism seems to show that the existence of epistemic subjects is necessary, as long as there is an intrinsic, direct statement of some necessary fact, be it a tautology or anything else. Hence the postulate of correlationism and the transfer principle are incompatible.

Modality in Science and Fundamental Modality

One way to escape this conclusion is to hold that the transfer principle cannot be applied to statements about contingency or necessity. If we think that contingency and necessity in science have nothing to do with contingency and necessity in the intrinsic sense, the above problem can be avoided.

However, this solution seems to lead to an excessive conclusion. Consider, for example, the following scientific statement.

Touching an electric wire carrying a high-voltage current carries a risk of death

This statement is a proposition about possibility. If the transfer principle is not applied to statements about contingency or necessity, then from this statement we cannot draw the inference

If I touch this high-voltage line (the one I am now looking at), there is a risk of death

Worse still, as long as there is some intrinsically meaningful proposition of necessity, it follows from the postulate of correlationism that the existence of epistemic subjects is necessary. I would not die even if I touched a high-voltage line. This position is that of what Meillassoux calls the “subjective idealist” (ibid., p. 97).

This position can be criticized by slightly modifying the line of argument with which Meillassoux criticizes “speculative idealism” (ibid., p. 47). To put it in a word, absolute idealism, or speculative idealism, gives rise to the problem of the possibility of the “incarnation” of transcendental consciousness. Here “incarnation” means the following. According to correlationism, the world is constituted by a subject correlated with things. On the other hand, however, the subject appears in the world as something that has a body, and the relations the subject has with things are constrained by the body. The “body” here stands in a certain relation, if not a complete one, to the (material) body that is an object of science. Scientific statements can be thought to have some consequences for this constraint by the body. But if scientific possibility differs from direct, intrinsic possibility, then scientific facts about the body become unrelated to events in the subject.

For example, consider the following medical statement.

If glaucoma develops, the visual field is likely to narrow

However, if the transfer principle cannot be applied to statements about possibility, then from this we cannot derive

If glaucoma develops in the body (the object described by medicine) of the “I” (as a conscious subject), then the visual field of the “I” (as a conscious subject) may narrow

This seems to be an excessive conclusion.

The Weak Postulate of Correlation In order to avoid the problem raised by ancestral statements, one might also consider weakening the postulate of correlation. Here I would like to examine the argument made by Christian Beyer [3]. Beyer summarizes Husserl’s position as follows.

the existence of a contingent object A requires the necessary co-existence of a subject either acquiring knowledge'' regarding Aor having the ability to do so’’ (Hua XXXVI, pp.~139f). This is nothing but ``the thesis of transcendental idealism: A nature without co-existing subjects of possible experience regarding it is unthinkable; possible subjects of experience are not enough’’ (Hua XXXVI, p.~156). ibid.

By contrast, Beyer considers a “principle of weak correlation.”

merely require what might be called real higher-order possibilities—possibilities for acquiring epistemic dispositions in counterfactual (or actual) cases where epistemic subjects would be co-existing—that may remain unactualized but could be actualized by someone properly taking into account a multitude of individual epistemic perspectives, by means of intersubjective experience. ibid.

According to the principle of weak correlation, for a fact P to hold, it is not necessary that there exist an epistemic subject who actually cognizes it; it suffices that it be possible, through appropriate intersubjective experience, to suppose the existence of an epistemic subject who cognizes that fact.

However, I want to point out that this principle, too, has the problem of ancestrality. The first problem is that supposing epistemic subjects in, for example, a universe that exists for only one millisecond, or a universe in which there is no matter at all, seems entirely impossible by means of ordinary intersubjective experience. Nevertheless, these are scientifically conceivable situations. Here I am not relying on the correctness of the physical theories that predict these situations. My ground is that excluding such physical theories because epistemic subjects cannot be supposed in them does not seem scientifically appropriate.

And, relatedly, if it is possible for epistemic subjects to exist in worlds far removed from our everyday world, such epistemic subjects are not, at least, “incarnated” in bodies of the kind we suppose. That epistemic subjects “incarnated” in such a variety of utterly different bodies are conceivable means that the epistemic subject is a “Pure Ego,” not subject to the constraints of a body. This, too, seems to give rise to the problem of “incarnation” raised in the previous section.

Summary In this essay, I have proposed a certain interpretation of the “problem of ancestrality” raised by Meillassoux. On this interpretation, Meillassoux’s argument consists of the “postulate of correlationism,” which implies that the existence of epistemic subjects is necessary, and the “transfer principle,” which implies that the existence of epistemic subjects is contingent. From the fact that these two are incompatible, I criticized correlationism, which interprets science fictionalistically. Further, I examined the assumption that modal propositions are not subject to the transfer principle, as a way of avoiding this problem, and showed that this approach cannot explain the conditions for the incarnation of consciousness. I also showed that attempts to weaken the postulate of correlation have a similar problem.

References

1. Quentin Meillassoux, Après la finitude: Essai sur la nécessité de la contingence. Japanese translation: Yūgensei no ato de (有限性のあとで, After Finitude), Jimbun Shoin, 2016. Page numbers cited for [1] refer to this Japanese edition.
2. Tsuyoshi Kurata, Lectures on Contemporary Ontology II (現代存在論講義 II), Shin’yōsha, 2017
3. Edmund Husserl in The Stanford Encyclopedia of Philosophy, Christian Beyer, Metaphysics Research Lab, Stanford University, 2018