> And:
> >What I actually think is that rules of the form
> >
> > If X has syntactic property A then X has phonological property B
> ## E.g. If a verb's subject has the syntactic property of being
> PRO then it has no phonology?
Yes, or rather, simply "If X is PRO then X has no phonology". Cf.
"If X is CAT then X has phonology /kat/".
> >are OK, but rules of the form
> >
> > If X has phonological property B then X has syntactic property A
> >
> >are not.
> ## I take it that you're assuming PRO is an example of this but I
> don't see why.
I don't have the necessary books here to check, but IIRC, GB had a
load of theorizing about syntax of empty categories, where empty
categories were defined by their null phonology. To me that is on
a par with defining a class of words that are monosyllabic and
start with a labial obstruent and then theorizing about their syntax.
I have just realized that I've been making a slightly different point
from the one I originally made & I'd forgotten it was the original
one that you were responding to. About the phonologylessness of PRO
being suspicious if accidental and conceptually clumsy if principled.
As I've subsequently said, I would accept it if a decent case were
made for there being a rule that words with certain syntactic
properties (simply "being PRO" is insufficiently explanatory!) had
no phonology. And if the phonologylessness of PRO is claimed to be
accidental, I can accept that there is some functional motivation
for this, but in such a case one would expect a lot of cross-linguistic
variation in the phonology of PRO; one wouldn't expect it to be
universally null.
My (very possibly wrong) impression of recent TG is that the
phonologylessness
of PRO is not accidental, but that it is not adequately explained. It's also
my impression that there is held to be a theoretically significant kind of
phonologylessness, evident in pronominals/anaphors, and an insignificant
kind
evident in various null prepositions and so on that get proposed here and
there.
--And.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|