Re: Pronouns in Split Ergative systems

On Mon, 13 Sep 2004 11:52:28 +0100, Chris Bates
<chris.maths_student@...> wrote:

> Describing Morphosyntax says that ergative marking in split ergative
> systems is always applied to less animate entities, whereas more animate
> entities take accusative marking.
Yeah. That's one kind of split ergative marking.

[snip]
> Are there any natural languages with
> a split ergative system divided like this where the split also occurs in
> the pronoun system, with pronouns referring to humans taking accusative
> marking, and pronouns referring to non humans taking ergative marking?
Funnily enough, I was just reading Dixon on that very subject (ergativity
split by sematics of NPs) earlier today. There's a hell of a lot to
excerpt, but the short answer is yes, the split can occur anywhere in the
heirarchy:

(Most likely to be Nom/Acc)
1st person pronouns
2nd person pronouns
3rd person pronouns
Demonstratives
Proper nouns
Human
Animate
Inanimate
(Most likely to be Erg/Abs)

Note that Accusativity and Ergativity can spread each from their opposite
ends of the spectrum by differing amounts. You might have a system where
there are a middle few of the above that take *no* Accusative or Ergative
marking (i.e. you must use word order or some other feature to
disambiguate), or where a middle few might have a full tripartite system,
marking S, A and P each with their own marker.

Some examples...

Dyirbal:

   1st & 2nd | 3rd
A -0        | -Ngu
S -0        | -0
O -na       | -0
   ACC       | ERG

Cashinawa:

   1st & 2nd | 3rd pronoun | 3rd noun
A -0        | habu~       | (nasalisation)
S -0        | habu        | -0
O -a        | haa         | -0
   ACC       | TRIPART     | ERG

In Yidiny, the patterning is

1st and 2nd pronouns - Accusative
Human deictics and interrogative pronoun - Tripartite
Inanimate deictics, proper names, kin terms - Tripartite with optional
drop of O marking
Inanimate interrogative, common nouns and adjectives - Ergative

In Nabe"b, it's

1st person - Tripartite
2nd, 3rd - Ergative

There are exceptions to this universal, but they're few in number.
Occasionally personal names rank above (more accusative than)
demonstratives and there are a few languages where 2nd is above 1st.

In addition to these and other Noun-based splits, Dixon covers a lot of
detail on Verb-based splits, both syntactically- and
semantically-controlled (which he labels split-S and fluid-S
resepectively). I can try to summarise those sections if you're
interested. Note that it's possible to have verbs split in one way and
nouns split in another way within the same language.





Paul