I've recently been re-reading Langton's Kantian Humility, and pondering a problem connected with her discussion. It seems to me that properties must be identified by what they do. After all, we can come to know about them, and how could we do that except via their effects? Langton argues for a thesis along similar lines, and attributes this view to Kant as well, though she confines the thesis to phenomenal (in the Kantian sense) properties.
However, two Humean theses also strike me as plausible. First, I agree with Hume's rejection of substance. Things are just bundles of properties. Second, I accept a Humean account of causation; causes just are regularities in the occurrence of events.
Combining these views produces a tight circle. Properties are determined by their causal powers, and causes are determined by the pattern of occurrence of properties. Intuitively, this is not a happy situation. This is perhaps part of the reason for Langton's endorsement of Kantian things-in-themselves, with unknowable intrinsic properties; the intrinsic properties of the things in themselves determine their behavior, so they enable an escape from the circle. The perfectly natural properties of Lewis may also be a way out of the circle.
I do not find any of the escape routes plausible. I'm more inclined to think that, contrary to appearances, no escape is necessary. Causal determination is perhaps necessarily asymmetric, but it is clear that whatever form of determination is going on here is not causal. Logical determination is often symmetric. That may not be what the situation involves either; surely what's going on is some sort of metaphysical determination, which many at least would distinguish from logical determination. But whether it's distinct from logical determination or not, it isn't clear to me that it couldn't share logical determination's capacity to be symmetric.
There are, admittedly, other issues raised by simply embracing the circularity rather than trying to find an escape, but I will postpone them for future posts on this topic.
 It would also be possible to escape from the circle by supposing that we do know intrinsic properties in the case of conscious experience, in the form of Humean simple impressions, or in what the moderns call qualia. But of course such a thesis would be absurd.