Because I can interpret it as there being no such case when there exist ice and stone separatedly, while the "sink to the bottom" is still technically true, because it's where the stone is at the end of the experiment.
Eh, man, let's just drop. It's became just a matter of arguing what the question "really" says, which part should be taken literally, and which shouldn't, and we could do this forever. If anything I'm sure we can agree that the wording of the problem is not perfect.
Here, let me give you a riddle as well.
Imagine a gigantic guillotine(i.e.two blades, one slightly angled), spanning infinitely large distance from side to side.
It's cutting blade's angle is very small(let's say, infinitely small).
Now if the blade falls, the cutting point at the intersection of the two blades starts traveling alongside the guillotine's length.
Can it travel faster than the speed of light? Why it can, or why it cannot?
Now, imagine an eqally gigantic pair of scissors. As with all scissors, the farther you go alongside the cutting arms, the faster they move towards each other as you make a cut. If you start depressing the handles*(or whatever that not-cutting end is called in English), will there be a pair of points alongside the scissor arms that is traveling faster than the speed of light, and why there is or isn't?
*We are assuming that the material they're made of won't break under any pressure, and that you've got a hand large and strong enough to actually make use of the scissors.