DPs and Demonstratives

[Mirzayan Armik]

  I'm not sure how much this will help, but I thought I'd send just a note 
about the pitch-accent of these constructions in Lakhota as I've seen them 
in some of my analysis.

I can't say anything too conclusive because I was actually concentrating 
on the pitch accents on other parts of intonational phrase when I did 
some work on the subject.

The piece of data I was working with didn't have too many explicit noun 
phrases that had both the dem and the det, but I do remember running into 
just a couple of cases of [dem N det] and [N det dem].

The one Lakhota [dem N det] construction that I remember was:

    le* aNpe*tu ki osni*yelo
    this  day   the  it's cold
    dem   N     det

In this clause the "le" has a rather high and broad pitch peak followed by 
a downstepped peak on the 2nd syllable of the noun. It is a bit hard to 
figure out what the "ki" is doing as it is really short, compressed 
between the noun and the following verb. It seems to almost participate in 
somewhat of an upstep into the verb that comes next. In either case, it 
definitely seems that the ki here is accentually dependent on the material 
around it (as was mentioned by Rory and John). I have a confusion as to 
the actual alignment of the pitch contour with the segmental tier in the 
"ki" region, so I can't say more on it without going back and looking at 
more cases (and with different lenghts of nouns and so on, if I can find 
such examples in the other spoken samples).

As for the [N det dem], I didn't find many cases of this in the 
conversation I had recorded. I found a couple of cases, but the 
intonational pattern seems harder to sort out than the [dem N det] or 
simply the [N det] cases. I have one case with (suprisingly) the same 
noun:

     aNpe*tu ki le ,   chaNte*-washte*ya wache*kiyapi  ....
     day     the this    with good heart they pray
     N       det dem

In the opening phrase of this sentence the 2nd syllable of the noun 
"aNpe*tu" has a high pitch peak and one sees a downstepped, gradual 
falling pattern on the ki-le sequence. The whole phrase is pronounced 
together, so there is no pause between the N and the det and there is 
definitely no pause between the det and the dem. That is, ki-le is very 
glued together, and my ear tells me that there is a secondary accent on 
the "le", but if there is one it is a very slight accent (almost invisible 
on the pitch track, which shows a somewhat gradual falling contour over 
the whole ki-le sequence). The /e/ on the demonstrative is lengthened, but 
the speaker has definitely put a phrase boundary after the "le" (which 
I've indicated by the comma above), so I would need different cases, at 
least one without a phrase boundary there and possibly also different 
lengths and accent locations of the N, to say anything more.

Don't know if this illuminates anything, but I'm afraid that's all I have 
for now.

Armik

On Wed, 22 Feb 2006, Koontz John E wrote:

> On Tue, 21 Feb 2006, ROOD DAVID S wrote:
>> Well, dem-det before the noun is absolutely impossible in Lakota, which
>> is the only one of these languages I know even a little about.  But
>> then, the order we have is always det-dem anyway.  The evolution of the
>> Dhegiha articles must be quite different from that of the Lakhota ones.
>
> Belatedly chiming in, I think this is definitely the case, and it is
> certainly consistent with Bob's work deriving most of the Dhegiha articles
> from the positional verbs:  noun det dem => noun det dem positional (= NP
> + V) => noun dem=positional.  Perhaps it would make sense in a purely
> diachronic context to consider the Dhegiha articles as something like an
> obligatory accompaniment of a now missing noun-final definite article in
> the Dakota fashion.  They are conditioned by (concordial in definiteness)
> with this deleted element, and concordial in position/shape with the noun.
>
> If this is true, then Dhegiha N dem=det would be expected to match Dakotan
> N=det dem approximately in functionality, and Dhegiha dem N=det to match
> Dakotan dem N=det.  I think this is consistant with what you and Rory have
> both said, i.e., I think you are both treating the posposed dem forms as
> resumptive and/or appositive.
>
> I always thought of the OP N dem=det forms as less marked, but I think we
> established contrary to my expectations (and without actual statistics)
> that dem N and dem N=det are actually more common in the texts.  I believe
> it is possible for dem=det N to occur, or even things like dem=det N=det
> and N=det dem=det, especially in modern usage.  I have never seen the
> article (det) before the noun without a preceding dem (or pro, perhaps, in
> the case of e=) to depend upon.  Like the Dakota definite articles the
> Dhegiha definite articles are obligatorily enclitic, and the
> demonstratives are not.  I don't know about pauses and prolongations, but
> I am pretty sure that the demonstrative is always a new high pitch.  I
> think dem N has both components accented, too.  I'm not sure about
> anything like downstepping or other possible indications of phrase
> structure and whether there is any difference in the two cases.  I
> completely agree with Rory that the definite article is swallowed up by or
> accentually dependent upon the preceding element.  It is never a new high.
>
> I'm not sure if indefinite articles are enclitic to the preceding element
> or not.  They never follow a demonstrative, as far as I know.  Dhegiha
> lacks the elaborate partitive realis coding of indefinites in Dakotan.
> It does distinguish waN (singular) vs. duba ~ j^uba (plural, plural
> diminutive).  I'm not sure if duba is partitive as well as plural.
> There is a sort of "topicalizer" =de that seems to have some properties in
> common with Dakotan =c^ha, but I'm not sure if it still exists.
>
> I'm waiting for clarification and probably some corrections from Ardis and
> Catherine!
>