Okay, one last time. <– Famous last words. I went back to Watkins [and also considered James Kreines’ “Kant on the Laws of Nature” (2009) – i.e., James Kreines who has been kind enough to have several protracted conversations with me about Kant in the last little bit, as has my good friend Robert Lawrence – and Tobias Henschen’s “Kant on Causal Laws and Powers” (2014)], and I poured back over the CPR and the Prolegomena too, and I still can’t get a Kant who hasn’t rejected the CPR to be able to have causal powers that are not, in the end, simply an aspect of the transcendental category labelled “Of Causality and Dependence.” I can see that even the CPR Kant might be read as wanting non-transcendental causal powers a la Aristotle – and certainly that a later Kant can be – but I can’t see how he can get them. (Watkins, to be clear, explicitly ascribes to Kant a neo-Aristotelian account of what causality just is.)
Here is one key constraint on the CPR-defending Kant:
The neo-Aristotelian causal powers to the existence of which he is purportedly committed would have to operate within the realm of possible experience (though they would not necessarily have to be directly observable). I say this because Kant is clear, it seems to me, that the causation that Hume got wrong, the causation that is up for discussion in the discussion of what causation is, is not the ‘causation from freedom’ that he contrasts with the ‘causation according to nature’ at A531/B559. The topic is ‘causation according to nature’. And the question is what that is.
It follows from this constraint alone that if Kant does – as per Watkins – believe in the existence of neo-Aristotelian causal powers, i.e., powers had by objects of possible experience, the display of which powers being what causation is, then the causal powers in question cannot be ones that are located in a transcendental rather than empirical realm, or are borne by non-empirical entities. (Especially, in addition to not being the relevant type of causation, the freedom had by the noumenal self, were one to think of it as a causal power, cannot be counted as as the causal power of an object of possible experience.)
Here’s another constraint:
Aristotelian causal powers are oomphy. They are such that powerful things are capable of being active (as a competent English speaker would use the term, albeit not equal to motion as such). It follows from this constraint that they can’t be equated with the, or a, “reason-for” or “explanation-for.” To put it differently, neo-Aristotelian oomphy causal powers are not the same thing as the Principle of Sufficient Reason.
It follows from these two constraints combined that oomphy neo-Aristotelian causal powers also can’t be metaphysically equal to what Kant refers to as the “derivative a priori concept” termed ‘activity,’ which is said to be a “predicable” vis-a-vis the category of causality. [A82/B108] This because the a priori concept ‘activity’ is neither empirical nor itself active.
This leaves us with one last option, I think, if we want the CPR-endorsing Kant to be a neo-Aristotelian: just as Kant’s empirical realism somehow gets him phenomenal objects that are not metaphysically equal to sensations had by subjects (if one reads Kant in this realist, anti-phenomenalist way), so too – one might say – does his empirical realism get him empirically real neo-Aristotelian causal powers, the display of which is what causation is.
There are two problems with this option, though, that I simply can’t get past.
First, Kant is pretty clear in the CPR about what he thinks causation is. And it really doesn’t seem to be what Aristotle said. I can’t think of any place in the CPR where he says that Aristotle was basically right about what causation is, and that he, Kant, means to uphold that same general view. It would have been easy to have said that, there, if that really is the view that he had meant to put forward. So I can’t help but think that it matters that he didn’t.
Second, the difference between the robust materiality of phenomenal objects (if one reads Kant as not being a phenomenalist) and the causal powers that we are thinking somehow just are, indeed, had by robustly empirical entities — the difference is that Kant doesn’t tell us that the materiality of empirical objects obtains in virtue of the categories of the Understanding, as their oomphy, active causal powers would have to, if empirical objects were to have them, given Kant’s commitment not just to empirical realism but to transcendental idealism.
To be sure, it’s an open question where the materiality that distinguishes Kant from a phenomenalist (if one so distinguishes him) does come from, since what is “given” is only the fact of sensation. [Even if one thought that sensations occur in virtue of “things in themselves” (Kant’s belief in the existence of which, other than in the case of the noumenal self, there are reasons to doubt), it’s a hard view to defend since it’s in the nature of the Kantian case both (a) that sensations can’t be sensations of noumena; and (b) that noumena can’t be causes in what would seem to be the required, “according to nature” sense, viz., so as to be able to produce sensations.]
But the point for present purposes (i.e., setting aside the analogous issue of materiality) is that even at the empirical level, it looks as though with respect to causation things are precisely as a non-Aristotelian Kant seems to be saying they are: viz., causation obtains thanks to the relevant category of the Understanding, not thanks to the (neo-)Aristotelian properties of (neo-Aristotelian)substances. And this is so even if (assuming that one wants to read Kant as saying the following, which I think Henschen, for instance, does) — this is so even if it is the case that, thanks to the subsidiary or derivative predicable that is the concept of “activity,” we cannot but experience phenomenal objects as having oomphy causal powers.
What am I missing? About a Kant who wouldn’t have to reject the CPR, I mean.