Thread by @PhDniX

The Quranic reading traditions are a great source of information of Classical Arabic in the early Islamic period. For this reason it makes sense that Jonathan Owens in his "A Linguistic History of Arabic" draws upon it frequently, but also badly.

A thread with annoyances:

Owens (pg.40) cites al-Farrāʾ as saying ʿĀṣim and al-ʾAʿmaš have -uh (or -Vh) for the 3sg.m. form. But this is an oversimplification and misreading of the evidence.

The forms al-Farrāʾ mentions are yuʾaddi-h, nuwalli-h, ʾarji-h and yara-h (i.e. not -uh at all!).

These are all readings of Šuʿbah ʿan ʿĀṣim (al-Farrāʾ does not seem to have been familiar with Ḥafṣ' transmission of ʿĀṣim, except for yara-h, which in the canonical transmissions is not transmitted from ʿĀṣim at all (only Hišām), but Ibn Mujāhid reports it.

These forms are much more interesting than the purpose that Owens seek to exploit them for ("Classical Arabic just had dialectal -uh"): These -h instead of expected -hū/-hī forms occur for each final weak imperatives and jussives in at least one canonical tradition.

This is why reading traditions are so interesting: They often contain forms that are describable through the Arabic grammatical tradition, but show variation that is never described by this same tradition.

Owens oversimplification obscures and misrepresents this.

Discussing ʾAbū ʿAmr's al-ʾidġām al-kabīr, Owens says that al-Kisāʾī in the transmission of al-Dūrī had al-ʾidġām al-kabīr and that the Ṭayyibat al-Nashr says this. It does not.
Al-Dūrī is a transmitter of BOTH AA and AK, and obviously here as the transmitter of the former.

What drives me nuts is it is not look Owens could not have checked this. In 2006 there were recordings of these transmissions available.

(I also don't understand why he uses a modern commentary on the didactic poem based on the Nashr, rather than, you know, just use the Nashr).

Neither al-Kisāʾī nor ʾAbū ʿAmr use ʾimālah 'regularly', and certainly not the type of ʾimālah Owens discusses (which is ʾimālah conditioned by the presence of kasrah and blocked by the presence of emphatic consonants. This is simply a misreading of Ibn Mujāhid.

What al-Kisāʾī *does* have is a phonemic distinction between /ē/ and /ā/, where the nouns and verb with a third root consonant y receive /ē/, and those with w receive /ā/. Thus: al-hudē/hadē al-ʿaṣā/daʿā.

This type of ʾimālah is very interesting:

Sībawayh does not actually describe this type of ʾimālah. To him verbs should always have ʾimālah regardless of final root consonant: al-hudē/hadē al-ʿaṣā/daʿē.

Others (like al-Farrāʾ) do recognise this ʾimālah and its an ancient retention of a contrast from Proto-Arabic.

I wrote a paper on this in @AENJournal :
arabianepigraphicnotes.org/journal/articl…

(where I don't actually mention al-Kisāʾī, which is honestly a little silly, but I was just getting into that material at the time).

ʾAbū ʿAmr has much more limited ʾimālah, bizarre to call it regular.

Now the type of i-umlaut ʾimālah Owens talks about does get transmitted for al-Kisāʾī, although not by Ibn Mujāhid, so he is accidentally right.

Ibn Mihrān's Ġāyah spends a whole section on Qutaybah's ʾimālah, which is irregular, but is close to Owens' ʾimālah when it appears

ʾAbū ʿAmr is, of course, not a Kufan reader. Owens gets this right elsewhere in the book.
More problematic is Owens' claim that all modern reciters today are apparently doing ʾimālah wrong. He suggests a diphthong [iə] rather than [e:].

Some modern dialects' results of ʾimālah are in fact diphthongs of this type, e.g. Eastern Libyan Arabic and Maltese, klieb (although in standard Maltese it has monophthongized again); but these are CLEARLY the result of breaking of /ē/, a crosslinguistically common phenomenon.

This is a well-known chain-shift phenomenon, described in great detail by the famous linguist Labov: Long vowels rise, and when they can rise no further they break into diphthongs. This happened in English, Dutch and clearly also in modern Arabic dialects.

I don't believe a well-established and talented dialectologist like Owens is not aware of this; he is simply being contrary for the sake of being contrary; the argument is also frustratingly weak: ʾimālah is marked with a kasrah before ʾalif, which to him means it spelled <iA>.

In a script that only has three vowel signs, one is at a loss how Owens expected it to be written otherwise. It is also painfully literal, and moreover put way too much trust in modern text editions. Text editions are NOT what the authors wrote in their own hand.

Just because editions of Sībawayh and Ibn Mujāhid both use a kasrah before ʾAlif to write the ʾimālah doesn't mean the authors originally did. This may very well have been the decision of the editor, and surely Owens doesn't believe ʾimālah is still pronounce /ē/ in recitation.

Also if we take Owens' argument based on orthography to its logical conclusion, we would have to conclude that say miʿzē also spelled kasrah مِعْزٖى in Sībawayh is to be read as **miʿzī??

Discussing hollow root ʾimālah Owens says Hamza (sic, Ḥamzah) reads ʾimālah on ten verbs whoe medial consonant is /y/.

He is wrong about this (ḫēfa/ḫawf = √ḫwf) and he is wrong about Sībawayh's rule. Sībawayh says the ʾimālah is applied whose 1sg. form is CiCtu.

Ibn Kaṯīr is *not* reported to recite ʾansā-nī-hū. ʾansā-nī-hu is a reading unique to Ḥafṣ.
Ibn Kaṯīr has invariably long -hū/-hī, but they harmonize! But Ibn Mujāhid is even more explicit than he needed to be and spells all of this out, unlike al-Dānī, to no avail.

While the readings are a rich source of linguistic data, it is important to take the material seriously. Qirāʾāt literature is specialized and difficult to grasp, there is no shame in not quite understanding the details right away.

But *all* it needs is enough humility to just sit down and take a moment to listen to a reciter in the transmission you're writing about. If the reciter does something different from what you expect, chances are it is you who is getting things wrong, not the reciter.

We have to conclude that Owens was neither knowledgeable enough to understand the material, nor humble enough to be corrected. Every single section that mentions the reading traditions in the book gets the facts wrong. That's an impressively terrible score.

You guys thought I was done?! Me too, but then I ran into this horrendously garbled section:

ʾAbū ʿAmr indeed has the option to do rawm and ʾišmām instead of idġām. But rawm isn't "labialization" and ʾišmām isn't "rounding and fronting". The example he gives is also wrong.

qāla rabbukum would *never* have rawm or ʾišmām. Even if one recites with either rawm or ʾišmām, qārʷrabbu-kum doesn't exist, it's always qārrabbukum.

Rawm is the replacement of a final vowel /u/or /i/ with ultrashort [ŭ] or [ĭ]. It cannot apply to qāla, since it ends on /a/.

ʾIšmām is not "rounding and fronting", it's a broad term "giving the 'taste' of something to something else', in this case implicitly ʾišmām al-ḍamm is meant: giving the taste of the /u/ vowel to a consonant, i.e. labialisation.

This can thus only be done if the vowel was /u/.

In fact, every single reader reads full assimilation with ʾišmām in one place: Q12:11 taʾmanʷnā < taʾmanu-nā

nquran.com/ar/ayacompare/…

Where he got the idea from that rawm is the labialisation is utterly unclear to me.

In other words the options of ʾAbū ʿAmr are for:
qāla lahum, yaqūlu lahū, bi-l-bāṭili li-yudḥiḍū
ʾidġām: qāllahum; yaqūllahū; bilbāṭilliyudḥiḍū
+ʾišmām: qāllahum; yaqūlʷlahū; bilbāṭilliyudḥiḍū
+ rawm: qāllahum; yaqūlŭ lahū; bilbāṭilĭ liyudḥiḍū

So while Owens says rawm *reduces* morphological marking it in fact increases it. With rawm, the three ending u/i/a are distinct (a has assimilation, and u and i are ultrashort).
ʾišmām likewise increases, adding an audible reflex of /u/.
Idġām reduces the most distinctions.