Print

Print


Sorry, I must call into some doubt the received wisdom of 
Wikipedia, etc. etc. and even the World Atlas of Language 
Structures (WALS) which you cited:
WALS http://wals.info/chapter/81

Much of the Wiki data on SOV vs. SVO is based on Russell 
S. Tomlin's book -- Basic Word Order: Functional Principles, 
1986.

Tomlin looked at 402 languages, 
41.8% came out SVO
44.8% came out SOV
9.2 % came out VSO

In fact, Tomlin subjected his own percentages to a 
chi-squared test and found that there was no significant 
superiority of SOV over SVO in the corpus. 

The WALS list seems to be 1377 languages
Subject-object-verb (SOV)	565
Subject-verb-object (SVO)	488
Verb-subject-object (VSO)	95

What comprises the corpus is of course the key.

WALS has, for example, only 10 forms of Chinese,
not even those that are far more distinct from each other 
than the many Turkic languages/dialects it does list. 
Does not have Gan, Xiang, Jin, etc. For Wu it has only 
Wu(Changzhou) but no Shanghainese, etc., etc. Perhaps
20 to 30 languages are lost in China.  Note how vacant 
the WALS map is there -- in the midst of those billions. 

Yet there are many, many Australian languages in the
WALS corpus, many forms of Northwest Caucasian 
languages, etc.

There are over 250 Austronesian languages. About 
200 Australian languages. Each of those listings 
account for more languages than the entire
Indo-European family of languages.

In general, I like the World Atlas of Language 
Structures: WALS http://wals.info/chapter/81
     ( But not that odd Figure 1.)

We must recall that WALS is the summary of a lot of 
Ph.D. dissertations and other now available studies 
in linguistics, so it reflects an uneven corpus. There's
lots of Soviet era works on 'the minorities,' and
lots of U.S. works on 'Indian languages.' Many  
dissertations done in other places easily gotten to by  
scholars -- Australia, Pacific Islands, etc. 

I see no evidence of anything systematic about 
language selection for counting. If there were
a thousand more Australian languages well 
documented they would be put into the list as
well.

So I wouldn't put too much faith in the idea that 
SOV is all that more typical than SVO. 

I feel it is safe to say that SOV is surely NOT 
'dominant in most places.' 

Best regards,              LEO

     Leo Moser  
-----Original Message-----
From: Constructed Languages List [mailto:[log in to unmask]] 
On Behalf Of Guilherme Santos
Sent: Wednesday, September 17, 2014 3:37 PM
To: [log in to unmask]
Subject: Re: Terminology for Sentence Structure

As you can see in WALS http://wals.info/chapter/81, SOV is 
indeed the most common word order. And can be seen in 
this feature (http://wals.info/feature/81A) that it is 
dominant in most places.

As for being the 'original' word order i can't really say anything

2014-09-17 19:05 GMT-03:00 Leo Moser <[log in to unmask]>:

> Over the years, I have seen many references to SOV being the original 
> or "the most common word order" in natlangs.
>
> So for many the "most common" or the "usual" word order begins with 
> the SOV supposition. Since it differs from Eng./ Sp./ Rus./ Chin./ 
> Fr./ etc., it appeals to conlangers who want something different. Yet 
> if it is "most common" order can it also be exotic?
>
> But is SOV "the most common word order" in natlangs.?
>
> I have several books by linguists that seem to document SOV.
> But I notice that the languages in the samples are often very minor, 
> endangered languages. In some studies large numbers of very related 
> (and minor) languages/dialects get into the count. (Doctoral 
> dissertations by linguists.)
>
> Might the "original pattern" have been no pattern at all?
> Look at Jiwarli.
> http://www.hrelp.org/aboutus/staff/peter_austin/AustinHale.pdf
>
>  Leo Moser
>
> -----Original Message-----
> From: Constructed Languages List [mailto:[log in to unmask]] 
> On Behalf Of Jeffrey Brown
> Sent: Wednesday, September 17, 2014 12:04 PM
> To: [log in to unmask]
> Subject: Re: Terminology for Sentence Structure
>
> Why not call it the "most common" or the "usual" word order?
>
>
> On Wed, Sep 17, 2014 at 8:20 AM, David McCann 
> <[log in to unmask]>
> wrote:
>
> > On Wed, 17 Sep 2014 07:52:16 -0600
> > Scotto Hlad <[log in to unmask]> wrote:
> >
> > > Thank you for your replies everyone. My conlang Asirka is 
> > > topic-prominent. My concern is the term for what could still be a 
> > > preferred structure of a sentence. It's been 37 years since I 
> > > studied Latin but I believe that you will most commonly see a SOV 
> > > structure in a simple declarative sentence. Is there a name for 
> > > this common structure? Scotto
> >
> > Unmarked would do — there's no "official name".
> >
> > Latin certainly has unmarked orders: SOV, NG, DemN, AdvAdj. An 
> > example of rhetorical order would be "Nobis non satisfacit ipse Demosthenes."
> > (Cicero)
> >
> > I remember reading an article by a Russian linguist about his little 
> > daughter's parsing of sentences. For some time after learning to use 
> > case endings properly, she still ignored them in other people's 
> > sentences, assuming that the first noun was the subject of the 
> > sentence and the second the object.
> >
>