source of GL (was BL accent patterns)

[Rory Larson]

David, thanks for your comments.  I hope the idea you describe isn't in the past tense, because it makes very good sense to me.  In Omaha, we do have /dh/ from epenthetic */y/ in some contexts, but not in that one.  There, it's still just epenthetic /y/, which we haven't been recognizing phonemically.

Lakota:	kaghA			kichaghA

Omaha: 	gaghE			gi(y)aghE

(I'm using /gh/ here for the voiced velar fricative, which you generally use g-with-a-hat for, and we have lately been writing /x/.)

The idea I proposed was that the intrusive "r" was an analogical development, while your proposal for *ki-ka- => *ki-ya- => *kya- => gla- is straight phonology.  If my idea is shot down, I would certainly favor yours.  In fact, I think the two might work together to reinforce each other.

I also enjoyed the German "r-emphatikum" discussion.  I never heard of that idea before!

Best,
Rory


-----Original Message-----
From: Siouan Linguistics [mailto:SIOUAN at listserv.unl.edu] On Behalf Of ROOD DAVID S
Sent: Friday, September 13, 2013 11:05 AM
To: SIOUAN at LISTSERV.UNL.EDU
Subject: Re: source of GL (was BL accent patterns)


Rory, I think your idea about an extra intrusive "r" is sort of like one I had a long time ago.  In the little paper I wrote for the Eric Hamp festschrift issue of IJAL I dealt with a weird Lakota fact (that is no longer true -- another casualty of bilingual Lakota speakers).  Older speakers always changed the dative of "kaga" 'make' into kichaga, with an aspirated ch, instead of the expected kicaga, which would be the normal palatalization of /k/ after /i/.  Since /ch/ in Lakota comes from PS *y, I proposed that at some point there must have been a rule that k disappeared after ki and the resulting kia sequence acquired an epenthetic /y/.  That epenthetic /y/ was salient enough to be included when other */y/ changed to /ch/.  I think I remember finding that there is a Dhegiha /dh/ from the same */y/ -- you can confirm that perhaps.  Anyway, your idea that an /r/ got in the picture isn't too far from my idea that we are dealing with a /y/ that appeared between vowels when a !
 /k/ got deleted.  If that's true, then we can say that ki > g happened before */y/ in the ka- verbs just as it did in the y- verbs.
 	I do not remember how we sort out the PSI difference between */y/ and */r/, but I do recall that they get mixed up sometimes.  Bob, we need you again.
 	Just as a little footnote to the /r/ idea: when I was in grad school I wrote a course paper about whether or not English "speak" is cognate with German "sprechen", given that there is no trace of an /r/ in the English.  I discovered that there are several sets of correspondences in Germanic with and without an /r/ in the initial or final cluster.  I remember the English doublet "wiggle' and "wriggle", and something about "spark", but the details are in a file cabinet in my office and I'm not there now.  I think German "Sprosse" got into the game, as well as some word for a thatched roof.  The 19th century writer who brought all of these together concocted a story about primitive folks around a campfire dealing with twigs and sparks and speaking noises that was quite amusing. 
What he proposed was a sporadic "r-emphatikum" that was used for some words and survived only in some of them in some languages.  Perhaps Siouan and Germanic both latched onto an "r-emphatikum".
 	What fun we can have with abstract phonology.
 	Best,
 	David


David S. Rood
Dept. of Linguistics
Univ. of Colorado
295 UCB
Boulder, CO 80309-0295
USA
rood at colorado.edu

On Fri, 13 Sep 2013, Rory Larson wrote:

> Jan,
>
> Thanks for that.  I’m glad to confirm that it works that way in Dakotan too.
>
> Sorry for the confusion about ‘vertitive’.  I was a bit fuzzy on that term, but googling and looking at a few entries just now I see the word is pretty much pegged to the concept of traveling back to an original point, which wouldn’t apply here.  The other two terms mentioned were ‘reflexive’ and ‘possessive’.  Would this be ‘possessive’, ‘look in at one’s own’, or ‘reflexive’, ‘look in at oneself’?  Is there actually a difference in the protocol of the language, either synchronically or diachronically?
>
> If the gla- does in fact originate from the combination of the prefix ki- with the instrumental ka-, then that blows away my analysis below.  If that is the case, I also hope that someone who understands how that works better than I do will correct me.
>
> Regards,
> Rory
>
>
> From: Siouan Linguistics [mailto:SIOUAN at listserv.unl.edu] On Behalf Of 
> Jan Ullrich
> Sent: Friday, September 13, 2013 2:56 AM
> To: SIOUAN at LISTSERV.UNL.EDU
> Subject: Re: source of GL (was BL accent patterns)
>
> Rory,
>
> If the proposed etymology of míyoglas’iŋ is correct then what is involved is a possessive, rather than a vertitive, I think.
> In possessive forms the instrumental prefix ka- becomes gla-, as in:
>
> kaksá ‘to cut sth’ --> glaksá ‘to cut one’s own’
> kahíŋta ‘to sweep sth’ –> glahíŋta ‘to sweep one’s own’
> okáštaŋ ‘to pour sth into’ --> ogláštaŋ ‘to pour one’s own into’
>
> This is why I think that oglás’iŋ (possessive) comes from okás’iŋ, although the former is not used as an independent lexical unit. It is not uncommon, however, that possessives (and other forms) of some verbs are used only in compounds.
>
> I think that historically the gla- form originates from the combination of the prefix ki- with the instrumental ka-. I hope that colleagues who have been working on the diachronic analyses will correct me if this is not the case.
>
> Jan
>
>
>
> Rory wrote:
>
>>> In this case, I think there has been an analogical replacement of a difficult *k-k- series that speakers didn’t want to hack their way through.  When they hit the first *k-, they were momentarily confused as to whether it was the k of the *ki- or the k of the *ka- their verb started with.  They opted for the latter.  Then they wanted to make it vertitive, and remembered from all their *kru- and *kra- and *kre- verbs that *kr- did just that.  The [a] that followed was the [a] of the *ka- rather than that of *ra-.  They just infixed an *r into the *ka- prefix to make it vertitive.  The resulting GLa- thus became vertitive for both *ra- and *ka-.  The difference is that for *ra-, it is the G makes it vertitive, while for *ka-, it is the L that signals vertitive/reflexive/possessive.
>
>>> Now if we could peep into something by means of our mouth, perhaps Lakhota would describe this with the verb **óyas’iŋ.  The vertitive form of this would be óglas’iŋ, and a mirror that we use our mouth to look into would be a miyóglas’iŋ.  This would have five syllables, counting the underlying *ki- that is represented by the G.  But the actual word is the homonym miyóglas’iŋ which is based on ókas’iŋ.  We do not have an underlying affixed *ki-.  We have only the instrumental prefix ka- which has been modified with an infixed L to signal vertitivity.  Assuming this analysis of the vertitive/reflexive/possessive of *ka- verbs is correct, I think the phonological argument would indicate that while miyóglas’iŋ, a mirror that we look into by means of our mouth, has five syllables, its homonym miyóglas’iŋ, a mirror that we look into forcefully, has only four.
>
>>> Hoping that Willem, Jan and David still support my etymology for 
>>> miyóglas’iŋ, and otherwise ducking and running for cover,
>
> Rory
>
>