Consider the vaguely Wittgensteinian sentence:
What cannot be said, cannot be thought.
I take it that this claim is open for debate and that it can be represented more formally as a universally quantified statement:
("p)(Ø(p can be said) ® Ø(p can be thought))
Now on a substitutional reading of the quantifier, this claim will be true iff for every substitution instance P of p the resulting proposition is true. Thus:
Ø(P can be said) ® Ø(P can be thought).
But it seems (am I missing something here?) that every such instance must be true since every substitution instance of the antecedent is self-defeating. Now if that is so, then (since it seems that the claim isn't, in this way, trivially true) that would give us some reason for thinking that unrestricted quantification in English may be objectual.
It might, however, be objected that the the antecedent isn't trivially false. Consider what happens if we allow, say, this book to be a substitution instance:
Ø(This book can be said) ® Ø(This book can be said)
One might claim that, even though This book can be said sounds ungrammatical in English, this is simply a contingent restriction on acceptable complements of the predicate said and that one can lift this restriction without thereby altering the meaning of the verb. If this is so, then it suggests that English has a covert method for implementing quantifier restrictions via (lexically specified) feature restrictions on verbs. In effect, despite the fact that the quantifier isn't explicitly a restricted quantifier, it nevertheless behaves as one as far as the semantics goes, namely, it behaves objectually rather than substitutionally. [Background assumption: restricted quantification in English is arguably objectual while unrestricted quantification is arguably substitutional.]