agreement conflict?

[Martin Haspelmath]

Dear Edith,

assuming that indexing by agreement actually involves something like
coindexation or coreference (e.g. in German _dem Vater(i) sein(i) Buch_
'the father's book', _dem Vater_ is coreferential with _sein_), the
Hungarian facts are not in the least surprising -- since there are no
two agreement markers in a sequence, we don't get 'Erzsebet's share of
the estate'.

However, explaining syntactic restrictions by morphological restrictions
works only if you believe that morphology is somehow God-given and needs
no explanation. If you also want to explain the morphology ("Why are
there no two nominal agreement slots in Hungarian?"), you have to say
something in addition.

It seems to me that the morphological restriction ultimately has a
syntactic explanation, because morphology serves syntax and comes from
syntax. And I believe that there are lots of dependent-marking languages
that don't allow two genitives. Thus, why can't you say "Pablo's
Corazon's picture" in English (meaning 'the picture depicting Corazon
that Pablo took')?

For some reason, grammars sometimes contain restrictions that seem to
limit expressibility -- a nice demonstration of the general point that
grammar is autonomous from function. Maybe a grammar that contains a
rigid single-genitive construction is in some way more user-friendly
(although it occasionally involves inconveniences for speakers), while a
grammar that allows multiple genitives is problematic (i.e. involves
other inconveniences). So I am wondering about the functional
explanation for the autonomy of grammar.

Martin

--
Dr. Martin Haspelmath (haspelmath at eva.mpg.de)
Max-Planck-Institut fuer evolutionaere Anthropologie, Inselstr. 22
D-04103 Leipzig (Tel. (MPI) +49-341-9952 307, (priv.) +49-341-980 1616)