Re: Doraja (was: Re: TRANS: a haiku)

Adam wrote:

>On Mon, 17 Apr 2000, Matt Pearson wrote:
>
>> Please describe how this new agreement morphology works.
>> What, for instance, do "ss" and "itr" stand for?  How does
>> agreement interact with argument fronting?  (I seem to recall
>> that Doraja is VSO, but with preverbal pronominal clitics and
>> rather free argument fronting.)
>>
>Glad you asked.  The basic story is this: Doraja verbs take an
>inflectional suffix which indicates: 1) for transitive verbs, the number
>and animacy of the subject and the number of the object; 2) for
>intransitive verbs, the number of the subject.
[interesting examples and discussion snipped]

I like your system a lot!  Does the introduction of agreement mean
that you no longer have 3rd person pronouns, or are these still
present?

>I'm still working out the implications of this entire agreement system.
>I like it -- it certainly frees up word order significantly, and gets
>rid of the whole _ui_ particle business that was causing me to wake up
>in a cold sweat; but I'm concerned that it might be unnaturalistic
>or perhaps even ambiguous in some cases.  Feedback is welcome.
Well, it could hardly be naturalistic *and* completely unambiguous.
Word order (and/or context) will still have to play a role in
transitive sentences where subject and object share the same
number/animacy features, and neither is topicalised.  But so
what?  As far as I can see, you've struck an elegant compromise
between free word order and lack of ambiguity.  Short of adding
explicit subject and object case-marking to the noun phrases
themselves, you could hardly have done much better.  I would
stick with it.

As for naturalism, my impression is that it's quite naturalistic.
I can't think of any natlangs that employ animacy agreement per se,
although the Algonquian languages come close.  The major difference
between Algonquian and Doraja is that in Algonquian languages it's
the animacy of the "absolutive" (intrans. subj., trans. obj.) which
matters for how agreement is marked.  As for marking number
agreement of objects on transitive verbs, this is hardly unusual.
Reminds me a bit of definiteness object marking in Hungarian.

The only thing that strikes me as even remotely unnaturalistic
about your system has to do with the difference in number of
agreement features between absolutives and ergatives (for want
of better terms).  My impression is that in languages where the
verb agrees in more features with one argument than with the
other, the argument which agrees for the largest number of
features will be the 'core' argument (viz. the nominative in a
nom/acc agreement system and the absolutive in an erg/abs
agreement system).  You have the opposite in Doraja:  The
agreement system is erg/abs, and yet the verb agrees with
more features for the ergative (animacy and number) than for
the absolutive (number alone).  This would make Doraja typologically
unusual, although I don't think it counts as an unnaturalistic feature.

In fact, it's not entirely clear whether Doraja is nom/acc or
erg/abs; it seems to have features of both systems.  The major
erg/abs property is the alignment of transitive objects and
intransitive subjects when it comes to agreement.  The major
nom/acc property is the passive/inverse marker "-(y)m",
which singles out topicalised transitive objects.

Making the system more consistently erg/abs would be quite
easy:  (1) Redefine the transitive agreement markers so that they
distinguish the animacy of the object rather than the subject, and
(2) recast "-(y)m" as a marker of topicalised transitive subjects.
Making the system consistently nom/acc would also be easy:
Just add animacy distinctions to the intransitive subject
agreement paradigm.  I'm not saying you should do either of
these things (I'm actually somewhat of a fan of typologically
muddled case/agreement systems), but since you solicited
suggestions, that's my two cents...

Matt.