Wolfgang Schwarz

Can you guess what kind of object the "totality of worlds" is, if it isn't a world?
Maybe this resolves the issue. (otherwise modal realism would be really nuts: each and every possibility is realized in a world & there are worlds for each and every 'possible' combination of realized possibilities?)

M.

I'm not sure I understand the question. I guess the totality of worlds is the mereological fusion of all worlds or the class of all worlds -- it doesn't really matter, does it? (The worlds are, as you say, such that each and every possibility is realized in a world & there are worlds for each and every 'possible' combination of realized possibilities. I don't think that's nuts, and I think it's not quite modal realism, for it leaves the nature of worlds unspecified.)

(a) The actual world, @, could not have failed to exist.
(b) So there is an actually existing thing, namely @, that could not have failed to exist.

First of all, @ is not an actually existing thing in the sense of 'merely actually'. That @ exists is true not just in @ but in all possible worlds. (So that @ exists is necessarily true, not merely actually true.)

Now, you also conclude "No actual thing could have failed to exist", from acceptable premises. I notice that the reasoning would extend also to possible things which are parts of non-actual possible worlds. For instance, talking donkeys. Take a specific talking donkey d from a world w, and you can run the argument, concluding that d could not have failed to exist. This is okay, in the sense that necessarily, there is a world such that d is a part of it.

Sloppily put, when our quantifiers also range over possibilia, 'could not have failed to exist' ceases to be synonymous with the predicate 'is a necessary being'.

OK! the last sentence of the post is really illuminating.
But "That @ exists is true not just in @ but in all possible worlds": I take it, that @ denotes the actual world, how come then that the existence of @ is true in @ and can be true in all other p.w.s?

M.

btw: sorry for dropping in, I do not know a thing about modal realism, but find it nice to play with this idea (as does everybody else, I guess)

Wo,

i guess there is a issue about context that crops up here. Suppose we have the claim that

"no actually existing thing could have failed to exist"

whether this sentence expresses a truth or a falsehood depends on whether you interpret the quantifiers as restricted or unrestricted. Unrestrictedly, everything exists necessarily.
Restricting our attention to various worlds, this is false, since there are worlds where i don't have counterparts, to take one example.

The question is what semantic contribution modal operators make to unrestricted claims. Here, i'm inclined to think that John Divers (1999) is correct. John thinks that modal operators, when they prefix unrestricted quantifiers are semantically redundant, endorsing the principle that Possibly P iff P in such cases. This is meant to work for what he calls "extraordinary" modal claims too (claims about things which are part of no world). Take "possibly there are properties". On a natural way of taking lewis, properties, qua sets, are not parts of worlds. So we cant take existence 'at' a world to be mereological in this case. So how should we take it if we are Lewisian? Again, Divers' idea is that we take existence 'at' a world as unrestricted existence, so the 'at w' operator is semantically redundant in such cases.

Isn't the problem here simply that you're using 'exist' in a sense that is inclusive of possible, but non-actual, existence? To pick up Irem's example, talking donkeys exist (in other possible worlds), they just don't exist *actually*.

If this is how we are understanding 'exist', then the conclusion seems entirely uncontroversial. It really just means 'no actually existing thing could have failed to exist _in some possible world_'. It does *not* follow that 'no actually existing thing could have failed to exist _actually_'.

Or am I missing something here?