Relative clause- case assignment

[Wolfgang Schulze]

Dear Colleagues,

when working with texts from Udi (a Southeast Caucasian language), I
came across a specific technqiue of assigning case to the head of a
relative clause (realized as a particple construction) which goes as
follows [note that Udi has ergative case marking)]:

Given that the head of a relative [participial] construction exerts
S-function (subjective with an intransitive matrix verb) and the
participle construction is overtly transitive (Objective is present),
the head (in S-function) is not marked by the absolutive (as it would be
expected), but by the ergative, cp.:

c^äli-g^-ox biq'-i adamar-en gölö irazi-ne-bak-sa
fish-pl-dat2 catch-part:past man-erg much happy-3sg:s-lv:intr-pres
'The man who has caught the fish is very happy'.

[we would expect '... adamar gölö irazinebaksa' if referring to the
valence of the intransitive verb in the matrix clause].

However, this kind of 'case attraction' does not alwas apply: The less
transitive the particpial construction is, the more likely it is that
the head reacts on the valence of the matrix verb, cp.:

c^äli biq'-al(a) adamar iräzio-ne
fish:abs catch-part:pres man:abs much happy-ref:abs-3sg:s
'A man who catches fish is a happy one'.

or [with so-called verba sentiendi (A>IO split):

ail-ux buq'-al(a) xinär (*xinär-a) ta-ne-c-e k'ua
child-dat2 love-part:pres girl:abs (*girl-dat1) go-3sg:s-past-perf home
'The girl who loves [its] child went home.'

The above-metioned technique is not allowed in case the head exerts
another function (such as the objective), cp::

t'eg^i c^äli-g^-ox bi'q-i adamar-ax (*adamar-en) te-zu be%g^-e
today fish-pl-dat2 catch-part:past man-dat2 (*man-erg) neg-1g:a see-perf
'Today, I haven't seen the man whoch has caught the fish.'

I know that some people probably relate this technique to what is
sometimes called 'relative clause internal head' constructions (in
formal approaches) [personally, I don't think that such a formal label
does explain anything]. In order to substantiate what I think about the
reasons behind the Udi technique, I would like to raise the question is
whether such a constructional type (together with its restriction to
heavy transitives in subordination) has good typological parallels. The
conditions are:

1. The case system of the language in question should be ergative (or
should have an ergative AGR-pattern);
2. The technique should be restricted to heads in S-function;
3. The technique should be restriucted to heavy transitivity in the
subordinated structure.

I gues that my query is rather trivial - most probably, I have simply
overlooked typological parallels or what has been discussed in current
treatizes of relative-like construnctions. But I would be happy if
people could help to fill my cognitive gaps (and gaps in memory).

Best wishes,
Wolfgang



--
********************
Prof. Dr. Wolfgang Schulze
Institut für Allgemeine und Indogermanische Sprachwissenschaft
Universität München - Geschwister-Scholl-Platz 1 - D-80539 München
Tel.: ++49-(0)89-2180 2486 (secretary) // ++49-(0)89-2180 5343 (office)
Fax:  ++49-(0)89-2180 5345
Email: 	W.Schulze at lrz.uni-muenchen.de
Web:   	http://www.lrz-muenchen.de/~wschulze/
********************