This is the final post summarizing some conclusions from Difficult Atheism, before this series launches out into new territory. In previous posts I have introduced the series, discussed a schema for distinguishing between different atheisms, sketched Alain Badiou’s interruption of the mytheme by the matheme and Jean-Luc Nancy’s “Christmas Projection”, and reflected upon Nancy’s own idea that there is “something in Christianity deeper than Christianity” itself. I now turn to Quentin Meillassoux and ask once more, in a preliminary way, whether there might be a moment in his thought that can be considered “theological”.
Meillassoux avoids both Badiou’s assertion of the unchangeable nature of philosophy and Nancy’s recourse to a Christian notion of the archetype in his “something in Christianity deeper than Christianity”. He does this by insisting that the only starting point for his philosophy is that there is no necessity. Note that this is his starting point, not his final position. We shall see below that a certain necessity does enter Meillassoux’s thinking (the necessity of contingency, and the necessity of the law of non-contradiction), but it is made necessary, precisely, by the need for there to be no necessary being or necessary law.
The radical nature of his position becomes apparent if we consider the import of this “only”. There are no eo ipso necessary laws, either of nature or of logic, and certainly no necessary being or beings. Why does Meillassoux insist on this starting point? Because allowing any necessity into philosophy would, in fact, be opening wide the door to religion. A belief in perennial laws is religious because it makes some transcendent action necessary in order to maintain the laws over time. Without such a metaphysical intervention there is nothing to guarantee that (natural or logical) laws may not change. Concomitantly, Meillassoux warns that ‘We have removed the gods, but we have kept the belief in the divine solidity of laws’ (L’Inexistence divine 4), reminiscent of Nietzsche’s “I am afraid we are not rid of God because we still have faith in grammar.”
For his own part, Meillassoux insists that these constants can be abolished, for the simple reason that nothing sustains them from the outside (ID 4). He deals at length with the obvious objections that could be raised to this adherence to the contingency of natural and logical laws, not least among which is the observation that natural laws have remained constant over a long period. We might resist Meillassoux’s notion that natural laws could change at any instant with the simple observation that they do not, in fact, change: Eppur non si muove! Meillassoux’s argument, in brief, hinges on the difference between chance (understood in terms of a finite and known number of possibilities, like a dice throw) and contingency (for which there is no known number of possible outcomes). Whereas chance presupposes a prior structure within which it operates (for example the structure of the faces of a die), contingency obeys no law and works within no such structure (ID 13). It follows, Meillassoux argues, that we cannot use probabilistic reasoning about the set of all possible worlds, because there is no set of all possible worlds (ID 36-38). Contingency is the appearance of a new universe of cases, not the appearance of any given universe (ID 16). We are therefore mistaken to refute Meillassoux’s thesis on the basis that the chance of a given law not having changed over a very long period of time (the argument that “if it can change, it would have changed by now”), because chance itself is only thinkable under a regime of the stability of physical laws, and so the objection assumes the stability it intends to prove.
Meillassoux builds his position as follows. First, there can be no real necessity, no necessary being, on pain of theology. Secondly it follows that the facticity of a thing is not itself a fact (Après la finitude 107/After Finitude 79), because if facticity were itself a fact (that is to say, contingent and not necessary) there could be a necessary being, and the door would once more be open to religious fideism. So, the only necessity is contingency itself:
what is, is factical, but that what is is factical, this itself cannot be a fact. Only the facticity of what is cannot be factical. Or again, in other words: it cannot be a fact that what is is a fact… The contingency of beings, and it alone, cannot be a contingent property of that being (ID 44).
Factiality, in other words, is the non-facticity of facticity (AF 107/AfF 79). Contingency is itself necessary in order to avoid a necessary being which, after the death of God, we have no grounds to admit into our thinking. We may say that an object is de facto red, but not that it is de facto de facto (ID 46).
And what is necessity? Necessity consists in the impossibility of qualifying contingency as contingent (ID 47). Contingency is necessarily non-contingent, because if it were contingent then there could be a necessary being, which Meillassoux has already ruled out. In order to avoid falling back into metaphysics, Meillassoux stresses that the principle of factiality does not maintain that contingency is necessary, but that only contingency is necessary (AF 108/AfF 80), as a direct correlate of the absence of any necessary being, event or law.
It follows from the principle of factiality that a radical change in the laws of nature, what Meillassoux calls a hyperchaotic change, is quite possible. Whereas mere chaos is ‘disorder, randomness, the eternal becoming of everything’, hyperchaos (surcontingence) is a contingency ‘so radical that even becoming, disorder, or randomness can be destroyed by it, and replaced by order, determinism, and fixity’, or again it is ‘the equal contingency of order and disorder, of becoming and sempiternity’ (‘Time Without Being’). Meillassoux evokes ‘a hyperchaos, for which nothing is or would seem to be impossible, not even the unthinkable’ (AF 87/AfF 64).
This hyperchaos itself issues in a residual necessity for Meillassoux because it is contingency which, unlimited and absolute, becomes its own norm (‘se norme elle-même’, ID 239 ; cf AF 90/AfF 66). Metaphysics announces ‘necessity is’, and relativism counters ‘there is no necessity’, but the principle of factiality stakes out a fresh position: there is necessity because necessity cannot be (‘il y a de la nécessité parce que la nécessité ne peut être’, ID 239 ; QM’s italics)).
Meillassoux’s ‘split rationality’: a theological moment?
What Meillassoux is trying to prove is his principle of factiality and the law of non-contradiction that he derives from it (a thing cannot be both itself and its opposite because then it would, after all, be necessary). In Après la finitude he makes it clear in a way that remained obscure in ‘L’Inexistence divine’ that the laws of logic are just as contingent as the laws of nature: ‘Everything could actually collapse: from trees to stars, from stars to laws, from physical laws to logical laws’ (AfF 53). Nevertheless, non-contradiction is derived from the principle of factiality and so stands, or so Meillassoux claims, as absolute, because it must be assumed in any attempt to contradict it.
But there is a problem with this, a problem I have called the ‘split rationality critique’ of Meillassoux’s proof, namely that he acknowledges that what is thought about (trees, stars, laws) is absolutely contingent, but he exempts from hyperchaotic contingency the thinking itself. The problem for Meillassoux is that, in order to be consistent, the laws of his own thinking should be subject to the same conditions as the natural and logical laws that thinking describes. In other words, the processes by which he arrives at the notions of ‘necessity’, ‘contingency’ and ‘factiality’ must themselves be able to be replaced by other, currently unimaginable, processes, and other ways of thinking.
In order to walk through this argument a little more slowly, let us consider Meillassoux’s reconstruction of facticity. In this reconstruction, he draws a necessity out of the strong correlationist model itself, where the strong model of correlationism is summarised as ‘it is unthinkable that the unthinkable is impossible’. In this correlationist model, the ‘logicity’ of the world does not conform to the structures of logical reason, and the givenness of the world in a representation does not conform to the structures of representative reason (AF 55/AfF 40). We cannot be sure that things are not Wholly Other to how they are represented to us. This uncertainty, the canonical limit of the rational, also legitimates faith in a God who transcends the limits of the thinkable (ID 49).
This reasoning yields a ‘precise and remarkable’ consequence: ‘it becomes rationally illegitimate to disqualify irrational discourses about the absolute on the pretext of their irrationality’ (AF 56/AfF 41).[i] Let us try to reconstruct in our turn what is at stake here. There are two instances of rationality in play in this quotation. First, the ‘non-rational discourse on the absolute’ which cannot be disqualified, and secondly the ‘rational illegitimacy’ of such a disqualification. It is the second of these two instances that shall detain us, because it is this second instance that Meillassoux fails to take into account in his argument for absolute contingency. In order to know whether or not it is rationally legitimate or illegitimate to suppose p, I must have some notion of rational legitimacy. But this notion of rational legitimacy, just as much as the ‘non rational discourse’ that is its object, must be contingent. Like gravity, it may be stable enough for the moment, but it is not necessary that it remain for ever thus.
So, to rephrase Meillassoux’s reconstruction of strong correlationism with this proviso inserted: it is rationally illegitimate, according to the contingent norms of rationality that prevail at the moment, to disqualify a non-rational discourse on the absolute on pretext of its irrationality. What would different norms of rationality look like? We do not know, and that’s the point. By thinking of them as “different” and as “norms” we are already projecting our current norms and our current understanding of difference onto them, domesticating hyperchaotic change.
This insertion of ‘according to the contingent norms of rationality that prevail at the moment’ is a modest proviso, perhaps, but one with the deepest consequences for Meillassoux’s principle of factiality. The very decision as to what may or may not be ‘rationally legitimate’ must not be unaccountably exempted from a possible future contingent rationality that in the present remains radically unforeseeable, on pain once more of fideism in the enduring necessity of rationality as it is currently understood and practised.
We cannot bootstrap rationality out of contingency, and to protect one’s own rationality from hyperchaotic contingency in this way is Meillassoux’s theological moment. Something is raised above and exempted from hyperchaotic change, namely Meillassoux’s own reasoning, and rationality is ‘split’ between, on the one hand, the object of thought (which is subject to hyperchaotic change), and the categories and reasoning of that thought, including concepts of ‘necessity’ and ‘contingency’ themselves (which appears not to be subject to that same change). Hyperchaotic change ceases at the cranium. Let Meillassoux be true and every man a liar; heaven and earth may pass away, but his words will never pass away (see Romans 3:4; Matthew 24:35).
 Hereafter: ID. All translations from ‘L’Inexistence divine’ are my own.
 Friedrich Nietzsche, The Anti-Christ, Ecce Homo, Twilight of the Idols and Other Writings, ed. Aaron Ridley and Judith Norman, trans. Judith Norman (New York: Cambridge University Press, 2005) 170.
 Hereafter: AF/AfF.
[i] ‘it becomes rationally illegitimate to disqualify irrational discourses about the absolute on the pretext of their irrationality’ (AfF 41).