It appears that I’m having trouble falling asleep this evening, which isn’t good as I have to be up early. In response to my last post, Tom of Grundledung was kind enough to remind me of a post I wrote nearly a year ago on the issue of normativity. On the one hand, I’m pleased by this post as it seems that my thoughts have been fairly consistent on these issues since they last flared up. On the other hand, as I review Tom and Pete’s comments, I find myself even more perplexed and wonder if we aren’t just talking about entirely distinct issues (i.e., talking past one another).

In a passage I quote in the post from last year, Pete writes:

I think the best point that can be made here is that there is more to normativity than ethical normativity. There is at least also rational normativity, which is prior to, and a necessary condition of, anything like ethical normativity. I would claim that it is indeed impossible to coherently deny the force of rational normativity. Regardless of the specific content of the fundamental norms of rationality (though we could suggest, for instance, the obligation to divest oneself of incompatible commitments), one must acknowledge that if one is engaged in an argument, then one is bound by norms which determine how the argument should take place, and that they are the same norms that one’s interlocutor is bound by. To put it another way, one may at time make claims like ‘well, I just use the word ‘justice” differently from you, but one cannot claim ‘I just argue differently than you do’.

This fact testifies to the binding character of certain fundamental norms that we implicitly acknowledge insofar as we engage in discourse at all. Some, myself included, think that this provides the possibility of a foundational approach in philosophy, in which deontology is indeed prior to ontology, grounded in that which none of us can deny insofar as we want to say anything at all. Whether or not such fundamental deontology can be extended beyond the theoretical into the realm of the practical and thus the ethical (as discourse ethicists like Habermas and Apel have attempted) is another matter.

I find myself grumbling a bit at Pete’s these that argument is necessarily grounded in certain norms. While I share with Pete a commitment to the principle of non-contradiction and identity, I also believe that we should look to rhetoric and how real life arguments function when raising this sort of question. The rhetoricians, I think, would have a very different perspective on this issue. With that said, I’m willing to follow him here.

For me the problems emerge when Pete asserts that deontology (and again, is this a specific Kantian reference or is “deontology” being used in a broader sense with which I’m not familiar) is indeed prior to ontology. What exactly is being claimed or asserted here? Is Pete making the claim that certain normative commitments are prior to inquiry, or is Pete making the claim that normativity is prior to being. These two claims are very different and have very different implications. If the former, then I don’t think the speculative realist, of whatever stripe, really has much of a dispute with Pete. Such a thesis doesn’t, I would think, commit one to correlationism or undermine realism. The realist here, I think, can simply shrug his or her shoulders and say “sure, there are norms that govern inquiry.”

read on!

This goes back to a question I raised in comments in response to Tom in my last post (here and here). What exactly are we referring to when we refer to “firsts” (Pete’s “prior to” relation). In a post I wrote a year ago on different kinds of ἀρχή, I noted that Aristotle speaks of ἀρχή or “firsts” in three different ways. There I wrote:

Aristotle distinguished between three different types of ἀρχή or principles, two of which, I think have largely been lost in the last 300 hundred years. When we speak of ἀρχή, we are speaking of the “whence” of things. According to Aristotle there are three different ways of speaking about the whence of things: ἔστιν, γίγνεται, and γίγνώσκεται. ἔστιν refers to the principle whence something is, γίγνεται refers to the principle whence something becomes, and γίγνώσκεται refers to the principle whence something is known.

In the context of these discussions, it is incredibly important to be clear as to whether we’re speaking about ἀρχή in the sense of ἔστιν (whence something is) or whether we’re speaking about ἀρχή in the sense of γίγνώσκεται (whence something is known). The first sort of principle is purely ontological in the sense that it does not refer to cognitive entities, propositional states, etc., but to what is first in being itself. The second sort of principle refers to mind and how mind comes to know being. First philosophy can take on very different senses depending on which sense of ἀρχή one is referring to. In the former case, ontology is first philosophy if one takes philosophy to be primarily about fundamental existence. Here questions of epistemology pertain to regional ontology or issues pertaining to particular beings, namely those capable of knowing. By contrast, if one takes first philosophy to be primarily a discourse pertaining to how we know rather than what is, epistemology becomes first philosophy.

The point here is that Pete and I might very well be talking about apples and oranges without realizing it. It sounds to me that Pete thinks that philosophy is primarily a meditation on knowledge, whereas I take philosophy to be a discourse primarily about being. While I do believe questions of epistemology have their place, I 1) don’t think epistemology can be characterized as first philosophy, and 2) believe that questions of epistemology and questions of ontology are entirely distinct. I say this with the caveat or reassurance that of course it is sometimes appropriate to raise the question of how someone can know this or that. For example, I do not endorse the idea that we should be able to rationally deduce the existence of God and God’s nature a priori in the manner of Leibniz or Spinoza. In fact, I don’t think we can rationally deduce the existence of any object. As a lot of us have repeatedly said, OOO does not tell us what entities exist, but only the most general features all entities have in common as entities. Moreover, I believe that I have repeatedly outlined the epistemological arguments that warrant me in these claims and have more or less outlined my theory of inquiry in detail, so I get a bit disgruntled that I keep getting asked this particular question about knowledge. As an additional caveat, I also agree with Pete that inquiry and argument should be governed by certain norms such as the principle of identity and non-contradiction.

In the post from last year, I quote Pete as further saying,

I think some of the suspicion concerning new forms of realism in continental philosophy arises from either its seeming lack of concern with these questions or its giving a nihilistic dismissal of them (e.g. Brassier — though I wish his book wasn’t so expensive so I could actually read all the details).

This is particularly pressing insofar as many people (like Asher Kay above) think that metanormative questions are prior to metaphysical ones. The thought here is that all inquiry implicitly assumes some standards which inquirers treat as demonstrating whether they are making progress. (This is, of course, not to say that they need a whole epistemology before they start: there does not have to be any ‘knowing before one knows’ as Hegel puts it.) I think the worry would be that the new realisms have not indicated how they will vindicate any such standards, even in the long run. The more eliminativist strands of SR in particular, with their sparser versions of materialism, may have trouble in doing this — in denying any robust ethical normativity, they may cut off the branch they are sitting on in respect of the theoretical norms required to articulate and assess their own position.

Previously, you’ve have set out what you’ve called a correlationist ethical approach, which joins a venerable tradition of ethical anti-realism. One question would be whether this is consistent with your ontology, where the propspects seems far rosier in light of the ontological pluralism of onticology. So, I think you’re in a more secure position than Brassier (and perhaps also Nick Srnicek) on this one.

Here’s where I find myself getting really skittish. On the one hand, while my own ontology is wildly at odds with Brassier’s ontology, I genuinely wonder how the conclusion is being reached that because Brassier doesn’t thematize or deal with these normative issues he rejects them. A far more plausible explanation would be that Brassier takes them for granted. And let’s be frank, at the risk of sounding anti-intellectual, isn’t it taken for granted and assumed that discourse should be governed by norms like the law of non-contradiction and the principle of identity? Here I’m just led to wonder where the mystery is.

In saying this, I’m not, however, being particularly fair to Pete, for he goes on to say that he’s worried we’re cutting off the branch we’re sitting on (i.e., these norms regulating inquiry and discourse) and that we need a way of grounding these norms. This is where I really start to itch and get suspicious. What, exactly, is being called for in this grounding here? In other words, what is the fine print. My nervousness here is that this call for grounding sounds suspiciously like the suggestion that we have to embrace something like a Kantian transcendental idealism or some variant thereof. And clearly, if this is the case, there are going to be real conflicts between the speculative realists and those that wish to ground these norms in this way.

With that said, I get the sense that there might be a fundamentally different reading of Kant at the root of disagreements between the speculative realists and the defenders of Kant. Readers familiar with this debate will recall Meillassoux’s argument from the arche-fossil which purports to point out problems at the heart of correlationism. Roughly the argument runs like this: Arche-fossils are traces of events anterior to the existence not only of humans but of life itself. Kant argues that time issues from the mind and applies only to appearances or phenomena, never things-in-themselves. Any claims about time as it pertains to things-in-themselves are, strictly speaking, meaningless. However, insofar as arche-fossils refer to a time prior to the existence of humans and therefore a time that must pertain to time not as a form imposed on the world by mind, but a time that pertains to things-in-themselves, we’re faced with a dilemma: Either the claims of the various scientists about arche-fossils are meaningless, or the correlationist account of time is mistaken. Meillassoux argument from the arche-fossil does not disprove correlationism, but purports to raise serious questions about the adequacy of the correlationist account of time. After all, Kant is attempting to ground science, so if his thesis about time forces us to abandon fundamental pillars of contemporary science on the grounds that it is dogmatic and idle speculation, this is a problem.

Now, having outlined this argument in detail over at Cogburn’s blog (including Meillassoux’s response to the standard rejoinder about conflating the transcendental), I was extremely surprised by the rejoinders I received from both Mikhail and Pete. If I understand his argument correctly, Mikhail argued that this argument doesn’t pose a problem to Kant because we are able to make inferences to these events based on the time that is accessible to us:

Me: “The whole problem, then, is that the arche-fossil presupposes the existence of a time anterior to the existence of the subject…”

No, it doesn’t. It does if you understand time to be absolute time (like Newton), but not if you understand time as Kant does. Arch-fossil is perceived by the subject in time and identified as a fossil. Remember all these questions I asked you about “big bang” and “inference” and so on?

Pete argued as follows (and in retrospect I think I glossed over his second paragraph a bit too quickly and completely missed it in my response to him):

There is no problem for Kant to account for the arche-fossil within time, any more than there is a problem for accounting for things that happened before he was born (e.g., Alexander the Great’s conquest of asia). This is because the correctness of judgments about objects is not indexed to *anyone’s* having a perceptual encounter with those objects, let alone me. Inference from observations I’ve had, and from observations that others have had (perhaps before I was born) are perfectly legitimate. The transcendental structure of representation is not somehow indexed to its manifestation in either any given individual or all individuals. The whole ‘before humans’ thing is a bit of a red herring.

This doesn’t mean that Meillassoux is wrong, because in response to this kind of defence he raises the issue of the manifestation of the transcendental, which is a pickle. I think this problem can be solved, but that it Kant be solved within Kant’s framework precisely because of his account of forms of intuition which are relative to our faculty of sensibility.

I am not at all trying to pick on either Pete or Mikhail here, but I was flummoxed by both of these responses. The problem with Mikhail’s response, as I see it, is precisely that Kant’s model of knowledge forbids the sort of inference Mikhail is evoking here. Why is this? Precisely because the arche-fossil refers to a “time before time”. Now when I say the arche-fossil refers to a time before time, I mean that it refers to a time anterior to transcendentally constituted time precisely because it refers to a time prior to humans and therefore prior to the transcendental subject. As a consequence, the arche-fossil necessarily refers to a time that pertains not to phenomena but to things-in-themselves. But this sort of time is precisely the sort of time Kant expressly forbids.

A similar problem emerges with Pete’s rejoinder. Meillassoux’s criticism is not the rather pedestrian claim that “we didn’t experience it, therefore we can’t know it.” Meillassoux is perfectly fine with inferences of the sort one might make when discovering an open bottle of mustard on the counter and concluding that someone made a sandwich, despite the fact that we didn’t observe the person making the sandwich. Meillassoux’s point is more fundamental: It’s that the arche-fossil refers to a time pertaining to things-in-themselves or prior to transcendentally constituted time.

When I pointed this out in my follow-up to Pete’s rejoinder, Pete wrote:

It doesn’t miss the argument. The question is whether it makes any sense to say that ‘time itself in Kant’s sense’ came into existence at any given time. Kant would deny this, and that’s why the form of the argument you’re presenting doesn’t work. If the correctness of judgments is not indexed to *anyone* then it’s not indexed to *everyone*. There’s no indexing to the emergence of humans within time any more than there’s an indexing to the emergence of me in time.

But again, I think this misses the whole point of Meillassoux’s argument: If Kant is right, time is indexed to the emergence of humanity (or some such rational being). Meillassoux is well aware that the defender of Kant will rejoin with the argument that he is confusing the transcendental with the empirical. Here the argument runs that the transcendental cannot properly be said to exist, but is instead a set of condition governing cognition and knowledge. Meillassoux’s retort is that the very idea of Kantian critique and the transcendental is premised on finitude (i.e., that we experience the world in terms of a perspective or a point of view) and that the condition for such finitude is attachment to a body. That body, in turn, is a real entity in the world that emerged at a particular point in natural history. Ergo, if we follow Kant’s reasoning all the way, we’re led to the conclusion that time, as he understands it, came into existence at a particular point in natural history. This is where all the paradoxes emerge, for this forces us into the dilemma of having to choose between Kant’s thesis that time pertains only to phenomena and never things-in-themselves, or a time that pertains to things-in-themselves or a “time before time”.

Now all of this seems tangential to the discussion of normativity with which I began this post, but actually it ties in. For me (and I don’t think I’m alone among the speculative realists in this interpretation of Kant), the key dispute with Kant revolves around his restriction of knowledge around phenomena or appearances, his claim that we can never know things-in-themselves, and his claim that mind structures reality. The point about phenomena is particularly important here. If this is the case in Kant, I simply don’t see how the Kantian can get around the challenge that Meillassoux presents with the arche-fossil because the arche-fossil requires us to speak of a time pertaining to things-in-themselves, not phenomena. It requires some form of real or absolute time.

In Pete’s and Mikhail’s rejoinders I sense, however, a very different reading of Kant, much more amenable to a realist position. The fact that Mikhail can argue that we can, for example, infer the big bang and that Pete argues that it makes no sense to talk about time coming into existence with the evolution of humans, suggests that they have abandoned or are ignoring Kant’s thesis about the restriction of time to phenomena such that time can never be said to pertain to things-in-themselves. This is the only way I can understand how they can simultaneously maintain that events like the big bang happened and that life and humans evolved and maintain Kant’s theory about the transcendental structure of things like time. To my ears, this would entail that Kant’s transcendental theory is far less radical than it seems. Rather than arguing that mind constitutes phenomena and we can never know whether these phenomena are anything like things-in-themselves, Kant’s transcendental theory would simply be outlining a set of norms or structures that pertain to reality itself. The argument would be something like “Look, all natural events occur in time and space, and the scientist must presume this in all their work. Moreover, all natural events fall under one of the twelve categories”, etc. Here Kant’s categories would be much closer to something like Aristotle’s categories (i.e., real structures of being), rather than structures of mind. This interpretation of Kant would explain why so many of the Kant defenders look at the speculative realists with bug eyed perplexity when we wax on about correlation. But if this is how Kant is being read, then I’m at a loss to understand how the transcendental dialectic can be understood insofar as the four antinomies are premised on showing how the forms of intuition and the categories of the understanding don’t pertain directly to things-in-themselves.