Members of phased kinds can cease to be members of that kind without ceasing to exist. Of course, as a matter of de dicto necessity, all toddlers are toddlers. But it’s not the case that all toddlers are necessarily toddlers. Members of individuative kinds belong to that kind as a matter de re necessity. A person can’t cease to be a person without ceasing to exist. A statue can’t cease to be a statue without ceasing to exist. If you flatten the statue, it’s not just that it ceases to be a statue; it ceases to exist.
Assuming you’re on board with those cases (i.e., assuming that your modal intuitions aren’t completely defective -- oh snap!), I’m wondering what you think about snowballs and islands. When you flatten a snowball, does it cease to exist, or does it just cease to be a snowball? Suppose you’ve got a low island, and it gets completely submerged every day at high tide. Does it cease to exist at high tide? Or does it continue to exist, but cease to be an island? Or perhaps you want to say that it still exists *and* it’s still an island at high tide?
I say the snowball survives flattening but ceases to be a snowball, and the island survives being submerged (and probably is still an island). If I’m right about this, it seems that Hawthorne’s complaint that (“barring an anti-realism that none of us should tolerate”) it’s intolerably arbitrary to include islands in one’s ontology and exclude (Hirsch’s) incars. One, but not the other, has extraordinarily weird persistence conditions. Similarly for Sosa’s claim that a realist who lets in snowballs is going to have to let in snowdiscalls as well. Anyway, my question is whether what I say about snowballs and islands seems right to you.