This is a good example of how the "natural" concepts are actually quite elaborate, paying utmost attention to tiny details that are almost invisible in other representations. But these details are in fact there, in the territory. The fact that they are small in one representation doesn't belittle their significance in another representation. And the fact that one object is placed in one high-level category and a "slightly" different object is placed in another category results from exactly these "tiny" differences. You can't visualize these differences in terms of quarks directly, but in terms of other high-level categories it is exactly what you are doing: keeping track of the tiny distinctions that are important to you for some reason.
That sounds right, but that sounds like I am (or at least could) visualize these levels as separate, since to keep track of the tiny differences that end up being important is impossible for my mind to do. This seems to necessitate that imagining irreducibility is not only possible, but natural (and perhaps unavoidable?).
This is not to say that irreducibility is logical, and our reason may insist to us that the painting is indeed reducible to quarks, whether or not we can imagine this reduction. But collapsing the levels is not the default position, a priori logically neccessary.
Here's our place to discuss Less Wrong topics that have not appeared in recent posts. Have fun building smaller brains inside of your brains (or not, as you please).