More on after all

Yesterday I noted this odd phenomena where after
all is usually used to split up a verb phrase headed by a cognate of to
be when the subject is a pronoun, rather than placing after
all in front of the verb phrase, but when the subject is a quantified
expression, this pattern is reversed. I thought this might indicate some
difference between quantificational and referential phrases, so I was
interested to see how the pattern transferred to definite descriptions. After a
bit more digging around, I found that (a) it’s hard to find definite
descriptions that are common enough to get any kind of usable data, but this is
unimportant because (b) names behave like quantified phrases rather than
pronouns, so whatever is going on here is not tracking a referential/quantificational
split. Some more data.

The only one of these that behaves like a
quantifier is Hussein. But even there
the data is messy, hence the asterix, because from eyeballing the data it looks
like 31 of the uses come from an advertisement that was placed in several
newspapers. So I have now gone from being puzzled by yesterday’s data to having
officially have no idea whatsoever about what could explain it. My tentative
hypothesis is that it’s to do with the length of the noun, and maybe its stress
patterns, but that doesn’t really explain why Bush and they are
so different.