Here is the logical argument:
[1] For all we know, A may be the state of some unknown part of B;
[2] C is determined by the state of some unknown part of B;
[3] Therefore, for all we know, C may be determined by A
It strikes me that premise [1] is extraordinarily weak. It says that A might be the state, so it follows that it might not be. It also says that it might be the state of an unknown part of B. But it might be the state of some unknown part of X, Y or Z, instead. Since nothing is known about this postulated unknown part of B (it's unknown, after all), then
any state might be the state of it, never mind A. Also, B might not even
have a part that A can be the state of, because you're only saying that A
might be the state. If it is, then we can deduce that B has an unknown part of which A is the state, but if A is not the state, then B needn't even have an unknown part.
There's also no guarantee that the "unknown part" referred to in [2] is the
same unknown part that is referenced in [1]. If it isn't, then A is irrelevant to C, quite apart from whether A determines some other unknown part of B.
Thus, the conclusion [3] is extraordinarily weak, too. Let's rephrase:
1. Let C be determined by the state, D, of some part, E, of B.
2. A might be a state.
3. A and D might be the same state.
4. A might be the state of some part, F, of B.
5. Parts E and F might be the same part of B.
We wish to draw the conclusion that C might be determined by A. This will be true if:
A turns out to be a state (2' doesn't help much there, and there's no premise in the original argument that says A is a state);
A turns out to be the state of part F of B (there's nothing in the original argument that tells us that it is).
E and F turn out to be the same part of B (there's nothing in the original argument that says they are).
A turns out to be the same state as D (there's nothing in the original argument that says it is).
That's a lot of "ifs".