The plane of logical syntax

In hist latest post on Quinean renditions of existential statements, Bill Vallicella argues that the so-called 'existential quantifier' of predicate calculus should be seen as a 'particular quantifier' acting against a background assumption of some set of existents.  Thus, '∃x.Cat(x)'  has to be understood as Of the things that there are, one of them is a cat.  The of the things that there are makes this assumption explicit.  Hence '∃' of itself, or rather the formula '∃x.Px', of itself, does not express existence, or at any rate, not the absolute notion of existence that Bill requires.  I think this is probably right, but it doesn't give Bill the result that he wants, because an 'ordinary language' understanding of existentials suffers from exactly the same problem. He says

If existence is to come into the picture, we have to get off the plane of mere logical syntax: there has to be some reference to the real world.

But this, I claim, is impossible, in both QuineSpeak and ordinary language.  The reason for this is that sentences must have meanings that are independent of the worlds in which they are said, though of course their truth values will vary across worlds.  This is why Some existing thing is a cat adds nothing to Something is a cat.  A fictional character cannot 'break out' of his fictional world and refer directly to the actual world merely by uttering the word existing.  If Frodo says Some existing thing is a hobbit he's talking about Middle Earth and he means this to be true.

We are stuck on the plane of mere logical syntax and we must settle for a relative understanding of exists.

Afterthought, Wednesday 22 August.

If I have got him right Bill is saying that the Quinean rendition requires the assumption there are some things in order for of the things that there are, one of them is a cat to make sense.  Well, suppose we allow ourselves to refer to the things with the caveat that there may be no such things at all.  Then one of the things is a cat allows us to infer there is at least one thing and hence there are some things. So the particular quantifier does after all carry a there are assertion.  Bill says

So to be perfectly clear, one must write:

Some existing thing is a cat. 

And now the explanatory circularity of the Quinean account is obvious.

But we have been here before in this discussion.  Bill is accused of introducing existing spuriously in order to create the appearance of circularity.  I simply cannot see any difference in meaning between Some thing is a cat and Some existing thing is a cat.