« The odd connected ṣ, ḍ, ṭ, ẓ and k of the Palermo Quran and Arabe 329f | Main

06/16/2018

Comments

Jadhimah

Deletion of the high second vowel in fa’ila and fa’ula was common in the speech of East Arabian tribes such as Rabi’ah and is found in some modern Najdi dialects. For instance in those dialects one would say sam’a instead of sami’a (in modern Najdi sam’).

In a book I have (in Arabic) on Rabi’ah’s dialect, it also mentions that the geminated form was raddit (ردِّت) instead of radadtu. This seems awfully similar to and possibly a precursor to widespread modern forms like raddayt/raddēt.

The Quranic form seems to stem from the same ancestral form CaCCt, and it seems this was common to Old Hijazi, and my guess is that while Quranic Arabic resolves the cluster by degemination, leading to CaCt, some of the spoken varieties of Old Hijazi inserted an epenthetic vowel leading to CaCCit, and eventually CaCCayt which we first see in Middle Arabic.

PhoeniX

Thanks for these interesting comments.

In fact, note that it is absolutely possible that Quranic Arabic actually had ẓallt or even ẓallit. The QCT does not allow us to be sure at all.

One can easily imagine how:

*ẓanantu would yield ẓanant

But:
*ẓaliltu would yield ẓall(i)t

It is, by the way, not just Rabi`ah that has forms like this. Also in the the Jbala region and Tafilalt region of Morocco do we find such forms.

I need to look at this more closely, but it seems that Andalusi Arabic exactly retains the distribution as we find it in Quranic Arabic!

Jadhimah

Yeah, actually what I failed to articulate in my comment is that the Quranic rule looks very similar to the Rabi’ah/modern Najdi rule for 3rd person non-double radical verbs (katab vs. sam’).

On the other hand the Rabi’ah rule for 1st person double radicals seems to differ from the Hijazi one since it operates on the [a] vowel as well. So you have radadtu > raddit in Rabi’ah, which would be illegal in Hijaz. Rabi’ah rule seems to be more analogical leveling of the radd- stem from the 3rd person than vowel deletion.


Coop

Sup Phoenix, I've read a lot of you and Glen Gordons work, excellent stuff. I'm just wondering if Glen Gordon still writes anything. He hasn't posted anything on his blog.

PhoeniX

Wow Coop, that is a very long time ago. I honestly have no idea what Glen Gordon does these days, no. Sorry!

Coop

I know this is a lot to ask but do you have any resources on Khoisan languages. I’ve searched the deep, dark recesses of the internet and all I found was Starostins Khoisan database (it’s a hell pit). Specifically I’m looking for one a grammar or dictionary of one of the ǃXuun or Ju languages.

Just to demonstrate how much of a hell pit Starostins database is, he actually thinks that his Proto-Ju word for cow, *gumi ~ *gome, comes from his Proto-Bushman word for “see”, *ɳǀV. Now I know why his ideas have such a bad rep. See http://starling.rinet.ru/cgi-bin/response.cgi?root=config&morpho=0&basename=%5Cdata%5Cbush%5Ckhoiet&first=1&off=&text_proto=&method_proto=substring&ic_proto=on&text_meaning=see&method_meaning=substring&ic_meaning=on&text_bsh=&method_bsh=substring&ic_bsh=on&text_ckh=&method_ckh=substring&ic_ckh=on&text_san=&method_san=substring&ic_san=on&text_had=&method_had=substring&ic_had=on&text_notes=&method_notes=substring&ic_notes=on&text_any=&method_any=substring&sort=proto&ic_any=on

PhoeniX

I don't know anything at all about Khoisan, so wouldn't be able to tell you. But I'll ask a colleague if she would be to provide some information.

I agree that Starostin's databases are usually not very useful...

HG

I heard you are looking for Khoisan resources! Specifically on !Xuun/Ju:
Heine, Bernd & Christa König. 2015. The !Xun Language. A Dialect Grammar of Northern Khoisan (Research in Khoisan Studies 33 33). Köln: Rüdiger Köppe.
Dickens, Patrick. 2005. A Concise Grammar of Ju|'hoan (Quellen Zur Khoisan-Forschung). Köln: Rüdiger Köppe.
Dickens, Patrick. 1994. English-Ju|'hoan, Ju|'hoan-English Dictionary. Köln: Rüdiger Köppe.

There are many useful contributions on all Khoisan languages in:
Vossen, Rainer (ed.) 2013. The Khoesan languages. London: Routledge.

Older work by Snyman can also be useful:
Snyman, Jan W. 1975. Žu|'hõasi Fonologie & Woordeboek. Kaapstad/Rotterdam: A.A. Balkema.
Snyman, Jan W. 1997. "A preliminary classification of the !Xũũ and Žu|’hõasi dialects". Namibian Languages: Reports and Papers ed. by W.H.G. Haacke & E.D. Elderkin, 21-106. Köln: Rüdiger Köppe.


Coop

Appreciate it HG, I'd be lost without your help.

Amine Gherensi

Going through this blog, I have noticed some posts about Berber and even a discussion about Latin words being loaned into Proto-Berber. Thus I was wondering if the information you have gathered could give further insight to what African Romance would have been like.

Coop

Phoenix, I recently posted a little something on Reddit about semivowels and diphthongs in PIE. I could really use someone like yourself to review it. Are and of my arguments valid for instance?
Here's the link, https://www.reddit.com/user/makutamakaveli/comments/9fyfk4/protoindoeuropean_semivowels/

PhoeniX

Did my blog get posted somewhere that suddenly new readers start posting with question on an unrelated blogpost or something? :-D

(don't mind at all, just striking to get two of these comments in a single day)

First Amine: Yes! Definitely, combined with place names that have an African Romance origin it should be possible to recover quite a few things.

We learn, for example that the nominative -us appears to be retained, e.g. a-fəllus 'chick' < *pullus, with the typical Romance semantic shift of 'young animal' to 'young chicken', but without the loss of the *s that yields pollo etc.

The accusative/neuter nom/acc -um appears to have lost the final nasal already: ɣasru 'castle' < *castrum.

Besides that, it seems that there was palatalization already of the velars, but it clearly had not yet developed into full-blow affricated forms of the type ʧ yet.

The Latin C is borrowed with a *k (but not the palatal *ḱ !) before front vowels whereas it is borrowed with (presumably still a uvular [q] at the time) before non-front vowels:

i-kikər 'chick pea' < cicer
ta-ɣawsa 'thing' < causa

Coop: I think you make excellent points. I don't think the argumentation why we should think of the *i and *u as purely vocalic counterparts to their consonantal value is very compelling. Nor do I think the question is very interesting.

Your idea that maybe stressed *i or *u were simply diphthongized to create the "full grades" *ei and *eu is an original suggestion. I think the more economical solution is still to assume that *i and *u simply behaves as any other consonant. Your suggestion of diphthongization does not significantly help understand the morphology of Indo-European any better. But it should be something to look out for when exploring deeper relations such as, for example, Indo-Uralic.

One thing your suggested solution does not solve, which the 'consonantal approach' does solve is the difference between *e and *o-grades of the *i/*u diphthong.

If we really start out with, say, *prus- 'to freeze' in its present stem but *prus- 'to freeze' in the perfect, how does Germanic end up with *preus- for the former and *prous- for the latter?

Coop

Update, the link to my post is now https://www.reddit.com/r/linguistics/comments/9gbr7p/semivowels_in_indoeuroean/.

Anyway thank you so much Phoenix. As for the difference between *e and *o-grades of the *i/*u diphthong, I figure that they reflected some type verb/noun, active/inactive dichotomy. Look at *h₂ówis, "the one who is clothed (in wool)" with a detransitive or middle/passive meaning marked by *o. Pulleyblank said something almost exactly like this.

I wasn't aware that there was any *prous- by the way.

Coop

If you think the diphthongization idea is original, you should check out one of my other theories of which I am not very proud. But hey, I just had to get it off my chest so I posted it on Reddit. I don't necessarily believe the theory, and I think it could easily be refuted.

Well, here goes, https://www.reddit.com/r/linguistics/comments/9gcl1g/my_theory_on_the_palatovelars/

PhoeniX

Well preus-/prous- is what lies at the heart of the English freeze/froze Dutch vriezen/vroor German frieren/fror.

The thing is: If you think that the *e and *o-grades have a functional difference, even in stems with *i and *u, I would be inclined to say that that is evidence that they were fully consonantal. There is no appreciable difference between: *preus/*prous/*prus versus *perḱ/*porḱ/*prḱ.

I will check out your theory on the palatovelars later. :-)

Coop

You're right, it's the biggest chink in my theory. I must admit that before I came up with the functional idea I was stomped. Maybe I'm wrong and PIE was like Abkhaz or Circassian, with only 2 vowels and plenty of syllabic resonants. Or if I am right, it might be some type of ancient vowel harmony-

*skid-je-ti (full-grade ye-present) > *skéid-je-ti (full-grade je-present)

*ḱi-mos > ḱói-mos

Even this probably has so many exceptions so as to render the idea terminated.

The second biggest chink is syllabic r and l. They defiantly aren't original but come from syncope, so it would follow that syllabic i and u also come from syncope. I can't find anyway around this, and crazy old Szemerenyi's insistence that they don't pattern the same isn't helping. Sdeath!

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)