I’ve been revisiting the argument against correlationism in chapter 1 of After Finitude. This has always struck me as perhaps the least interesting part of the book, and certainly the most misleading, insofar as it tends to draw the focus away from the more principled overcoming of correlationism (by pitching it against what Meillassoux now calls “subjectalism”) in chapter 2.
Nonetheless, the fact that I’m now back to working on Meillassoux draws me back into the fray. And it seems important to grasp what the proper strength of the arche-fossil argument is, to attain the clearest and strongest formulation of its challenge to correlationism. I guess part of what’s drawing me back is also the assessment of Meillassoux’s arguments adumbrated in comments here: Despite having what I think is a decent grasp of the nuances of Kant’s position, I find Meillassoux’s argument to be quite a piercing critique of the Kantian framework, and it’s worth trying to articulate why. In fact, this post will turn out to be a defense of the cogency of the correlationist position, but I’m far from sure whether the rejoinder works, and it is the first real response to Meillassoux’s argument I’ve managed to come up with, several years after first reading it.
As far as I can see, the very core of the argument turns on the transcendental, more precisely, the transcendental subject as something that must be ‘taking place’, in pp. 24-26 of After Finitude. Here, Meillassoux introduces the distinction between the transcendental subject as being “exemplified” and as being “instantiated.” This distinction corresponds, if I understand correctly, to the old distinction between universals ante rem, the Platonic Forms, and universals in rebus, the Aristotelian forms. Does the transcendental subject exist over and above, and prior to, its instances, i.e., ante rem, so that its instances are merely exemplifications? Meillassoux claims, and rightly so, that correlationists must answer no to this question, to avoid being dogmatists. So the transcendental subject exists only in its instances, i.e., in rebus, it exists only insofar as it has empirical instantiations. But this means that one can talk about (and investigate scientifically) the emergence of the transcendental subject, its passage from non-existence to existence (since one can talk about and investigate the emergence of its first instantiation), and importantly, one can talk about this passage as something that takes place in time (since one can talk about the passage from non-existence to existence of its first instantiation as something that takes place in time).
What is important here is that Meillassoux is not as it were imposing a foreign conception of time onto the Kantian framework, not as the argument has been presented up until now at any rate (I will return later to whether he eventually tries to do that). Kant has to agree that the first instantiation of transcendental subjecthood is something that took place at a certain point in time – at a certain point in time precisely as Kant understands time, that is. Transcendental subjecthood does need to be instantiated, and what instantiates it are empirical subjects with a certain temporal duration, and this instantiation did happen for the first time at some point (given Kant’s hypothesis concerning the origin of the solar system, it’s pretty clear that he thinks there was a time when the earth was around but there were no humans on it). But here comes the problem: Kant is also committed to claiming that time is nothing other than the form of inner sensibility of this transcendental subject. Thus, one might reasonably think that if the transcendental subject needs to be instantiated, then the form of inner sensibility of this transcendental subject also needs to be instantiated, and that these instantiations are co-extensive (abstracting for now from the fact that Kant thinks it is logically possible that a transcendental subject could have a different form of sensibility). So the first instantiation of the transcendental subject is also the first instantiation of the form of inner sensibility, i.e., time. But – paradox! – the first instantiation of time is now shown to be something that takes place in time. Time emerges in time. Kant cannot claim that time is ante rem, somehow existing over and above particular subjects with sensibilities; that would obviously reinstate the time-as-thing-in-itself conception. So he seems forced to give up the idea that there was a time prior to the first instantiation of transcendental subjecthood. A hefty price to pay.
The crucial question, I think, is this:
Can the in rebus instantiation of a universal be a condition of the possibility of the very thing in which it is first instantiated?
I think that the answer to this question is “yes,” because it seems that the instantiation can be what makes the thing in which it is first instantiated the thing that it is. So, presumably, the Aristotelian form (say, the form of a specific animal species, e.g., giraffehood) was thought to be a condition of the possibility of the individual things, i.e., the individual giraffes, in which it was instantiated, because they wouldn’t be giraffes if they didn’t instantiate giraffehood. This is the case whenever the universal in question is part of the essence of the thing in which it is instantiated. Thus, in many cases the instantiation of the universal is not the condition of the possibility of the thing, since the property instantiated is inessential to the thing being the thing it is. E.g., I instantiate “tiredness” now, but I will (presumably) still be me after drinking my cups of coffee tomorrow morning, despite the fact that I will no longer be instantiating tiredness.
Let us look at how this question pertains to the matter at hand, by considering some particular instances:
1. The in rebus instantiation of transcendental subjecthood is the condition of the possibility of the empirical subject in which it is first instantiated.
First of all, note that this is definitely true within the Kantian framework. And moreover, it seems to follow the same pattern as with the Aristotelian form considered above: The empirical subject would not be the thing that it is, i.e., an empirical subject, if it were not an instantiation of transcendental subjecthood. Being an instantiation of transcendental subjecthood seems to be part of the essence of being an empirical subject.
2. The in rebus instantiation of transcendental subjecthood is a condition of the possibility of the time in which it is first instantiated.
This case is slightly different. Time is not a thing in which transcendental subjecthood can be instantiated. It is, if one follows Kant, a form of intuition. Nonetheless, one can still say that time is something “in which” or “within which” transcendental subjecthood is first instantiated. And again, the claim is definitely true within the Kantian framework. No transcendental subject, no time; but at the same time this transcendental subject must be instantiated in time: “it is all the same whether I say that this whole time is in Me, as an individual unity, or that I am to be found with numerical identity, in all of this time” (Critique of Pure Reason, A363). Why is this? Again, the justification for the claim seems to be that the entity in which the universal is instantiated would not be the entity it is if the universal was not instantiated in it: Time would not be what it is if transcendental subjecthood was not instantiated in it, in other words, transcendental subjecthood is part of the essence of time.
At some point along this line of argument, Meillassoux will have to protest. And I think the protest must come with respect to sentence 2. One way of protesting would be to ask what the argument for the claim “transcendental subjecthood is part of the essence of time” is; why on earth would one think such a thing? (Kant’s main arguments are in the Transcendental Aesthetic and the Transcendental Deduction). But this is not Meillassoux’s approach, since he wants to argue not that there are insufficient grounds for asserting the Kantian thesis, but that there are sufficient grounds for rejecting it, along what I think are the following lines:
“2 is false, because it can be shown (and is shown, by science) that transcendental subjecthood cannot be part of the essence of time. It simply cannot be the case that time would not be the thing it is, i.e., would not be time, if transcendental subjecthood was not instantiated in it. For science tells us about a time before the instantiation of transcendental subjecthood, from which we can conclude that time is capable of being what it is, i.e., time, without instantiating transcendental subjecthood, since time, as it was then, was the thing/entity it was without instantiating transcendental subjecthood.”
But here comes the main point I want to try out in this post, however inchoately: Does this challenge by Meillassoux not presuppose something concerning what time is? Consider the following answer:
“Time is not a thing which moves in time, thus one cannot speak of time “as it was then” or “time as it is now.” If transcendental subjecthood is instantiated in time, it is instantiated in time, period. It is true that other objects can begin instantiating, or stop instantiating, certain universals. E.g., I can instantiate the universal “youthfulness” now, but stop doing so in, say, a year or so… But time is not that kind of object, as can be seen by the following consideration: If one were to talk about “time as it was then” and “time as it is now,” one would have to say that time changes, but changes can only take place in time, thus there would have to be a second time, time2, in which these changes in the first time, time1, could occur. So one should talk about changes taking place in time, rather than talking about time itself as changing.”
Is this a trick? Could one rephrase the point using, e.g., “parts of time” instead? If the instantiation of transcendental subjecthood is supposed to be the condition of the possibility of every part of time as well, then it might be claimed: “But a part of time, e.g., the part of time stretching from 16 to 15 million years before now, was what it was, e.g., a part of time, even though transcendental subjecthood was not instantiated in it. So transcendental subjecthood cannot be part of the essence of parts of time, and therefore cannot be part of the essence of time.” But I don’t think that this rephrased argument is sound. A universal can be part of the essence of an object, without being part of the essence of any part of that object. Let us stipulate that “philosopherhood” is part of my essence (doubtful though it is). Surely, that doesn’t commit me to claiming that any part of me has “philosopherhood” as part of its essence. So the fact that science can tell us about parts of time where transcendental subjecthood is not instantiated does not tell us that transcendental subjecthood is inessential to time considered as a whole. And since according to Kant, time as a whole is logically prior to its parts, one cannot claim that certain parts of time could happen independently of transcendental subjecthood being instantiated, even though they can certainly happen independently of transcendental subjecthood being instantiated in them.
I am still puzzling over whether this correlationist rejoinder can be adequate. And over whether the approach to the problem that I have considered here is on the mark: Is Meillassoux’s task really that of showing that “instantiating transcendental subjecthood” is not an essential property of time? Is the appeal to essential properties legitimate at all?
But if it works and is on the mark, I think that it dispels the worry of sliding from innocent-looking idealism to dinosaurs-are-God-testing-our-faith idealism faced with the ancestral realm: What it does, I think, is to position the ancestral back on the same footing as the rest of the empirical world. For it establishes that the instantiation of transcendental subjecthood can be a condition of the possibility of the time in which this instantiation arises, without it being necessary that transcendental subjecthod is instantiated in any particular part of time. So there is nothing particularly significant about the specific parts of time where transcendental subjecthood is, as a matter of fact, instantiated. What is significant is only that it is instantiated at some point of time or other. And this is how the transcendental subject can be correlated with all of time, but instantiated only in some of time.