Characterising the PSR
For Meillassoux the PSR is a philosophical fifth postulate. According to this principle (2008: 33), for “every thing, every fact, and every occurrence, there must be a reason why it is thus and so rather than otherwise.” Meillassoux implicitly equates the PSR with the view that “every entity is absolutely necessary” (ibid). It is an interesting question whether he would also accept the equation of the PSR with necessitarianism, the view that every truth is absolutely necessary. I don’t think it matters for present purposes. In any case, let’s add a bit more detail to what the PSR says before we continue.
A proposition p lacks sufficient reason if it is “essentially not deducible” (2011: 181). Thus we can think of sufficiency in terms of deducibility, and hence the elimination of possibilities: a reason q for p is sufficient only if it excludes all possibilities incompatible with p. Finally, a reason q cannot be its own sufficient reason if it is a contingent truth. Given this traditional understanding of the PSR, which I shall call “Spinozist”, there is a straightforward incompatibility between sufficient reasons and contingency: if p is contingently false in the actual world W, then there is no q sufficient for p in W. So, for any q that would entail p, q is either necessarily false (and hence the PSR false since this precludes a sufficient reason for p), or contingently true, in which case the algorithm iterates. Result: the set of all such truths produced by this regress is either contingently true (in which case it cannot be its own sufficient reason), or necessarily true (in which case everything deducible from it is also necessarily true). Either the PSR is false or there is no contingency.
Meillassoux’s project and its (contingent?) incompleteness
Meillassoux’s project is to carry out an “adventure” analogous to that of non-Euclidean geometry, this time within philosophy (2008: 92). But whereas Lobachevsky developed his hyperbolic geometry without first trying to demonstrate that the fifth postulate was false (i.e. without trying to demonstrate the consistency of Euclidean geometry with its negation), Meillassoux believes he can demonstrate that the PSR is false. Nevertheless, Meillassoux acknowledges that his project is incomplete in its present form. This acknowledgement comes in his discussion of the competing axiomatisations of set theory, which I do not discuss here. This is one sense in which After Finitude is, as Badiou says, a “fragment” of a larger philosophical enterprise. There is a second sense, however. Meillassoux wants to prove what he calls the “principle of factiality” or “unreason”, i.e. the falsity of the PSR and consequent reality of contingency. But what we are given in After Finitude is only a part of this proof. Whether it is the most important part or not is a moot point, since what is left out is also unquestionably important: namely a proper confrontation with Hegel and the subjectalists, who hold that some correlate necessarily exists. Instead Meillassoux joins a long line of continental anti-Hegelians who postpone confronting Hegel in print, sometimes indefinitely.
Correlationism is a hybrid of scepticism and relativism. Scepticism to look away from x and relativism to look towards y. (This is to be contrasted with non-sceptical fallibilism, where we keep looking towards x, only less assuredly.) According to the correlationist, the intrinsic properties of things cannot be distinguished from artefacts of our particular subjective mode of access to them. Meillassoux quotes Wolff’s gnomic expression of correlationism (2008: 6):
We are in consciousness or language as in a transparent cage. Everything is outside, yet it is impossible to get out.
The basic correlationist move is something like this: human knowledge is receptive, passive, finite etc., hence we are ignorant of how things are like independently of our access to them. If this is true then we can reason as follows: since the in-itself provides the sufficient reason of the given, it follows that the latter, like the former, is inaccessible. The reason may be there, but we can’t know it. This is what Meillassoux calls “facticity” (2008: 41). Meillassoux wants to reveal facticity as absolute, i.e. as factiality or real contingency. He claims to have a proof of factiality, but what does he mean by this?
A proof? The revenge of incompleteness
Meillassoux describes his proof as indirect and refutational, yet ecumenical and absolute (2008: 61, my italics):
One establishes the principle [of factiality] without deducing it, by demonstrating that anyone who contests it can do so only by presupposing it to be true, thereby refuting him or herself.
How seriously should we read “anyone” in this passage? Well, the way Meillassoux proceeds presupposes that self-refutation is unacceptable – thus radical sceptics are excluded straightaway. But never mind – the real question is whether all non-sceptical inquirers are subject to Meillassoux’s dictate here.
To illustrate what may be problematic about this, allow me to briefly contrast two understandings of facticity: minimal and substantive. The former corresponds to doubt per se, and the latter to anything beyond this. Plausibly there is a connection between the former and the correlationist themes of passivity, receptivity, and finitude: only we finite, non-divine underlings doubt. If that is enough for factiality then we can stop right here – our work is done. Whether right or wrong, I do think that this way of reading Meillassoux is heuristically productive. It casts light on aspects of his position that otherwise seem arbitrary, in particular the sheer reductiveness of grouping Berkeley, Hegel, Nietzsche, Deleuze and others together as members of the same philosophical family (namely subjectalism). Each of these thinkers denies the inference from doubt to brute contingency.
Of course we are tempted to add: rightly denies this inference, since (according to this temptation) it deserves little more than an incredulous stare. Surely minimal facticity is not enough to establish the falsity of the PSR! Indeed, it is typically thought to be insufficient even for establishing correlationism. Prima facie, only the most radical subjectivist could potentially encounter difficulty in accommodating minimal facticity in her worldview. At any rate, to show that e.g. Hegel suffers from the same difficulty would be just as hard as showing that his position collapses into radical subjectivism.
More seriously, I suspect that, not only is it just as hard to
(a) show that commitment to minimal facticity leads (without further commitment) to factiality
(b) convince Hegel to accept premises stronger than minimal facticity*;
but, in addition to this, it is also the case that (a) is impossible, or at least not finitely tractable. Again, what I am suggesting is that if we make the premises of Meillassoux’s proof weak and thus hard to deny, then the validity of the proof becomes dubious. On the other hand, if we make the premises strong so as to secure validity, they become easier to deny. My current hypothesis is that between Meillassoux’s conclusion and its negation there is a stubborn undecidability: there is no valid argument with premises that cannot be denied on pain of self-refutation by anyone (who accepts that self-refutation should be avoided). That is, Meillassoux’s project is not merely contingently but necessarily incomplete: the subjectalist wiggles free.
If I am wrong about this then the correlationist must have an argument against the subjectalist that is anhypothetical. What might this argument be? I explore this question in my next post.
*If you think this is accomplished via the so-called paradox of absolute knowledge, I’d be interested to hear why.
- Quentin Meillassoux, After Finitude: An Essay on the Necessity of Contingency, trans. Ray Brassier, (Continuum: 2008).
- Graham Harman, Quentin Meillassoux: Philosophy in the Making, (Edinburgh University Press: 2011).