From Peter van Inwagen, "McGinn on Existence" in Modes of Existence: Papers in Ontology and Philosophical Logic, eds. Bottani et al., Ontos Verlag, 2006, p. 106:
There is the theory of Quine, according to which the two oppositions [that between being and non-being and that between existence and non-existence] are not two but one. Existence and being are the same. Existence or being is what is expressed by phrases like 'there is,' 'there are,' and 'something is.' And, similarly, non-existence is what is expressed by phrases like 'there is no, 'there are,' and 'nothing is.' Thus, 'Universals exist' means neither more nor less than 'There are universals,' and the same goes for the pairs 'Carnivorous cows do not exist'/'Nothing is both carnivorous and a cow' and 'The planet Venus exists'/'Something is the planet Venus.' This outline constitutes the essence of Quine's philosophy of being and existence.
And an accurate and succinct outline it is. But it just reinforces me in my conviction of the wrongheadedness of Quine's version of the thin theory of existence.
I grant that existence and being are the same. My objections begin with the assimilation of 'exists' to 'something.' The following are logically equivalent:
Cats exist There are cats Something is a cat.
and the same goes for:
Mermaids do not exist There are no mermaids Nothing is a mermaid.
But the thin theorist goes beyond the relatively uncontroversial claim of logical equivalence to the eminently dubious claim that the meaning (van Inwagen uses this word above) of 'exist(s)' is exhausted by the meaning of 'something' and the meaning of 'not exist' is exhausted by the meaning of 'nothing.'
To sort this out, we first note that 'something' splits into 'some' and 'thing.' To appreciate this, observe that the following are nonsensical
Some is a cat Thing is a cat.
Equally nonsensical are their canonical counterparts:
(∃ )(x is a cat) ( x) (x is cat).
So both 'some' and 'thing' are needed for 'Something is a cat' -- '(∃x)(x is a cat)' -- to make sense.
Now it is obvious that existence is not expressed by 'some' or '∃' since these are merely signs for particular (as opposed to universal) logical quantity. Existence is not someness. Existence is not expressed by '∃.' And it is obvious that existence is not expressed by the variable 'x,' which is merely the canonical stand-in for the third-person singular pronoun, 'it.' It is obvious, I hope, that one cannot express the thought that cats exist by saying 'It is a cat.' Existence is not 'itness.' Existence is not expressed by 'x' any more than it is expressed by '∃.'
So existence cannot be expressed by the quantifier part of 'something' or the variable part. Is existence expressed by both together? No. Putting together two pieces of mere logical syntax just gves you more logical syntax. 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. Suppose we write 'Something is a cat' as
Some thing is a cat.
But now the cat is out of the bag. For surely these things one is quantifying over are existing things: 'thing' is a variable having existing values. 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. We were promised an account of existence in terms of the so-called existential quantifier. But the account on offer presupposes the very 'thing' we want an account of, namely, existence. Clearly, one must presuppose that the objects in the domain of quantification are existing objects if the logical equivalences above mentioned are to hold.