Re: a grammar sketch...

> From: Jörg Rhiemeier
>
> Yoon Ha Lee wrote:
> >
> > On Sat, 30 Sep 2000, Joerg?= Rhiemeier wrote:
> >
>   [...]
> >
> > > It indeed looks like an active case system, and there are
> unfortunately
> > > no standardized names for such cases as apparently, all
> active languages
> > > in the "real world" seem to be head marking.
> >
> > I'm a little lost.  What does "head marking" mean?  Is that the same as
> > when languages put adjectives after the noun, etc.?  Or have I gotten
> > the terminology reversed?
>
> Head marking has nothing to do with word order, it means that the
> semantic
> relations between verbs and nouns are not marked on the noun by cases,
> but by agreement on the verb.  However, I do not understand head marking
> deeply enough to explain properly.  Marcus, please?
Actually, I believe it is a more general term than this.  Head marking
includes any form of marking the relation between a head and a dependent
where the head is morphologically marked rather than the dependent.  The
head of a phrase is that element that determines the syntactic function of
the phrase. A noun is the head of a NP, a verb is the head of a VP, etc.
Head marking thus includes things like marking the possessed head rather
than the dependent possessor in a possessive construction.  Your example of
marking grammatical relations on the verb (head) rather than case marking on
the noun (dependent) is another example of head marking.

AFMCL, amman iar has an interesting mix of ergative dependent case marking
and active head marking. Thus,

\f Galdor slew the dragon.
\t i galdranne eleth an i feng ernurgoiraen
\n galdar ir feng orgorte

\t i   galdranne        eleth             en         i
\m i   galad =an   -e   el-        -eth   en         i
\g the galad =masc -[A] assertive- -past  agt.to.pat the
\p det nam   =gnd  -erg mood-      -tense ptp        det
\x the Galdor           did               agt.to.pat the

\t feng        ernurgoraen
\m feng   -0   er-  en-    ur-  coiro -ae      -n
\g dragon -[P] do-  cause- not- live  -agt/pat -actn/proc
\p n      -abs agt- caus-  neg- v     -val     -vc
\x dragon      slay

Here 'Galdor' (galdran -e) as the A-function argument takes the ergative
case (-e) while 'dragon' (feng -0) as the P-function argument takes the
absolutive case (-0).  This is classic ergative dependent marking driven by
the syntactic relations of the dependent arguments to the head verb.
However, the head verb 'slay' (ernurgor-ae-n) exhibits active head marking
by taking the agt/pat valency marking (-ae) which shows the semantic roles
of its dependent arguments.

David