LINGUIST List 10.1331

Fri Sep 10 1999

Disc: Universal Word Order

Editor for this issue: Karen Milligan <>


  1. Joseph F Foster, Re: 10.1320, Disc: Universal Word Order

Message 1: Re: 10.1320, Disc: Universal Word Order

Date: Thu, 09 Sep 1999 16:51:51 -0400
From: Joseph F Foster <>
Subject: Re: 10.1320, Disc: Universal Word Order

At 10:31 09/09/1999 -0400, Larry wrote:

>>All six logical orders (3! = 6) of Predicate/MajorArguments are actually
>found. Some examples:
>OSV: Apurin´┐Ż (Arawakan; Amazonas), Xavante (Gean; S. Amazonas),
>Djirbal (Australia; like Warlpiri, Djirbal doesn't really have fixed
>word order, but this is the most common order found when NP arguments
>are nominals)

While I have no information on the others listed and they may well be OSV,
I am apprehensive of including Dyirbal in this category. If we read
carefully the extensive "glottography" (ethnography~grammar) of R M W
Dixon, we find that in the ordinary unmarked sentence in Dy., the so called
O noun is actually the syntactic subject. In the Dy. equivalent sentence

 The girl kicked the ball and rolled down the hill.

it is not the girl but the ball that down the hill rolled. But the order
would be

 ball-nom girl-erg kick-Tns .... roll-Tns. 

To get the "subject" girl into "SO" order and nominative case, one must use
an antipassive, i.e. additional morphology on the verb. The transitive
ergative sentence is then, as you point out, the unmarked sentence in
Dyirbal, both discoursewise and morphologically. 

So Dyirbal is probably not insightfully classed as OSV. The semantic
patient is the subject in the normal declarative sentence. What then is the
"S" noun? Obviously not the 'object' unless our words are to cease to have
any content reference and be used only as labels. It is normally the
semantic agent and in an oblique case. We know there are many languages in
which it is not only possible but common to have subjectless sentences.
(And I am not willing to posit underlying PROs all of which must then be
"dropped" solely for the ideological purpose of claiming that all languages
are alike.) Maybe Dyirbal is better thought of as a language with
"objectless" sentences? 
 (For pronouns in 1 & 2 person, of course, Dyirbal normally exhibits
nominative~accusative morphology.) 

Joe Foster
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue