Dick:
> And:
> At 19:46 05/07/2002 +0100, you wrote:
> >We have been using reflexives as evidence for covert subjects.
> >E.g. "The lotion is difficult to cover myself with" would be taken
> >to imply a covert ME as subject of (at least) _cover_.
> >But now that I think about it, I can't see why there couldn't
> >just be some kind of lexically empty subject position which binds
> >_myself_, with it being _myself_ that forces the empty subject
> >position to be interpreted as referring to the speaker.
> >
> >Are there arguments/rationales that I am failing to see?
> >
> >If not, is there any way of telling where there is a covert subject
> >and where there is an empty subject?
> ## As usual, what you need is *syntactic* evidence, otherwise all you know
> is that there's a semantic role which is earmarked for the subject (if
> there is one).

Ah, right. So I should be looking for nonovert subjects referring to
scales, balances, oats, wheat.... Thanks.

> For example, you might argue (and you might be right) that
> YOU and ONE are synonyms in (1):
> (1)     One/You shouldn't chat in lectures.
> If they really are synonyms, then by definition they must have the same
> semantic representation, so the agreement in (2,3) must be syntactic:
> (2)     One shouldn't amuse oneself/*yourself in other people's lectures.
> (3)     You shouldn't amuse yourself/*oneself in other people's lectures.
> And it will follow that "amusing" must have a covert syntactic subject in (4).
> (4)     Amusing oneself in other people's lectures is wrong.

That's what I had been thinking. But then it had then struck me
that if (4) contained an empty subject position that binds _oneself_,
it would still be fine, because it would mean that the 'referent'
of the empty subj position satisfies the meaning of ONE.

It depends how you formulate the agreement rule, I guess. If it
positively requires agreement then (4) would be bad with an empty
subject. If it simply prohibits disagreement, then (4) should be okay
with an empty subject.

> On the other hand the agreement in (5) is clearly *not* syntactic:
> (5)     The committee kept itself busy.
> (6)     The committee kept themselves busy.

"The scales have worn themselves out"
"The balance has worn itself out"

"The winner will have worn themself out"

> So maybe the agreement in (2,3,4) isn't syntactic either, but
> semantic/stylistic - the badness of "one ... yourself ..." has exactly the
> same sociolinguistic explanation as the badness of (7), where there's
> clearly no syntactic agreement:
> (7)     One shouldn't say that one is/*you are tired if it's not true.
> In short, I don't know of any evidence in English for covert syntactic
> dependents - which is why Chet and I had to base our argument on Ancient
> Greek! That wasn't just a way of showing off that one of us knows it ...

Oh, I think there's a fair bit of English evidence for covertness; your
paper cites the VP ellipsis argument I put forward, after all.

It's the covert vs. empty distinction that I'm struggling with, but
at least you've shown me a way forward.

--And.