Actually I can argue, quite clearly and logically, that any given object of roughly homogeneous composition can be subdivided into a significantly larger population of units that share the same basic identity as the source.
In example: Your stone block is composed of, say granite. I then hit your block with a hammer, and break it into 10 identically shaped, equally sized fragments of identical mass. Each of the 10 pieces is still granite, is it not? So we could describe your 1 stone block as being composed of 10 fragment-units of a granite block. If your block were to have a mass of 100 kilodwarfs, we can just as easily say that it is a stack of 100, 1 kilodwarf units of granite.
...just assume that the stack is actually one rock, of a volume of "7 stones". Therefore a big rock ( 8 ) plus 2 pebbles ( 1 each ) is enough to make a door.
But in my post I was not speaking of rock "blocks", which are created objects. I was speaking about mined rock, ore, plants etc. There is an abstraction involved in the conversion of stones[10]->block[1], due to the possibility (actually likelihood) of the ten "stones" of material being from different sources.
But this is no greater abstraction than assuming that if a miner fails to succeed at a mining check, that all the stone in the tile magically vanishes. And that he does succeed, the stone comes out in one uniform lump of constant mass that is suitable (exactly, with no left over) to make 3 mugs or one door.
Neither makes any sense in real life. However, to use roughly your own words, realism can be suspended on occasion for the sake of gameplay.
But no, I can not argue that "a block of granite is composed of 10 smaller stones". I can only argue, as I have been all along, that the very same chunk of rock might contain 10 units of rock.
So I am going to concede the point, because obviously you care more about this minor detail than I do.