Leafing through al-Zamaḫšarī's al-Mufaṣṣal fī al-Naḥw today, I ran into his chapter on ʾibdāl "replacement", namely his section on the replacement with tāʾ of the consonants wāw, yāʾ, sīn, ṣād and bāʾ. This leads to interesting reflections on Arabic grammatical theory. 🧵
Most of Arabic morpho-phonological theory deals with a concept known as ʾaṣl "root, origin", which is an abstract underlying representation of a word. It has similarities both to a phonemic underlying form, and etymological origin, but is neither exactly.
Rather, it is more of a Platonic ideal representation of an underlying form. The 'source' form from which the surface form (or forms) can be derived through a set of rational rules (ideally). For example, the ʾaṣl of the verb qāla 'he said' is {QaWaLa}.
In the most abstract of modern theoretical phonological theories, one might argue for this as the underlying form, but most would consider the synchronic form to be /qāla/ or /qaAla/. Indeed, Zamaḫšarī tells us that the wāw and yāʾ undergo ʾibdāl to ʾalif in this situation:
"The replacement (to alif) of the two sisters (wāw and yāʾ) is regular, for example, for qāla {QaWaLa}, bāʿa {BaYaʕa}, daʿā {DaʕaWa}, ramā {RaMaYa}, bāb {BaWaB}, nāb {NaYaB} whenever (wāw & yāʾ) carry a vowel and there is a letter carrying a fatḥah before them."
In this same chapter he also discusses the replacement of wāw and yāʾ by tāʾ, citing examples like ittaʿada "come to an understanding" ittasara "become easy", the iftaʿala form of the verb waʿada "to promise", yasira "become easy", understanding this as {WTaʕaDa} and {YTaSaRa}.
The idealized underlying forms often turn out to be the identical to the proto-Arabic forms, as is the case with *qawala 'to say', and *ramaya 'to throw', and it can be tempting to understand ʾaṣl "origin" as "etymology". But the grammarians are very strictly non-historical.
This becomes clear in Zamaḫšarī's discussion of the "replacement" of ṣād and sīn by tāʾ as examples of this he cites ṭast with ʾaṣl {ṬaSS} "a kind of brass vessel", and liṣt {LiṢṢ} "robber". In both cases we are dealing with loanwords so we know the tāʾ is original.
liṣt "robber" is transparently from Greek λῃστής /lēistēs/ "robber". In the Old Arabic as found in the Safaitic inscriptions we also evidence for this original t-pronunciation (though with ṭāʾ, not tāʾ!).
ṭast seems to be borrowed Middle Persian tašt, still around in modern Persian طشت, تشت "bowl"

In other words, in both cases we are dealing with an ancient assimilation st, ṣt > ss, ṣṣ, yet Zamaḫšarī decided to treat the assimilated form as the ʾaṣl.
Theoretically this is a little surprising: a shift of ss, ṣṣ > st, ṣt is completely isolated within Arbaic. Other words that end in ss or ṣṣ do not 'replace' the second sīn and ṣād with tāʾ, so why did he opt for this analysis?
It is tempting to see a normative bent in Zamaḫšarī's choice here. Apparently he takes liṣṣ and ṭass as the normative pronunciation of these words, whereas liṣt and ṭast were more marginal (today liṣṣ is indeed standard, but ṭast/ṭašt won out instead).
A possible other or additional explanation for Zamaḫšarī preferring this rather ad hoc explanation is that assimilation of consonants in synchronic Arabic is regressive, i.e. the former consonant assimilated to the latter. Progressive assimilation is (almost?) unheard of.
In fact assimilation in the other direction, that is tṣ > ṣṣ indeed occurs in Quranic recitation, and is analysis as assimilation instead of "replacement". ʾAbū ʿAmr, Ibn ʿĀmir, Ḥamzah, al-Kisāʾī and Ḫalaf all read ḥaṣiraṣ ṣudūru-hum (for ḥaṣirat ṣudūru-hum) in Q4:90.
I hope this look into Arabic grammatical theory was interesting and helps you in reading Arabic grammatical works.
I want to stress just how amazing Zamaḫšarī's al-Mufaṣṣal is. It's succinc, and mostly a summary of Sībawayh's al-Kitāb but MUCH clearer. Recommended reading!
If you enjoyed this thread and want me to do more stuff like it, please consider buying me a coffee.
If you want to support me in a more integral way, you can become a patron on Patreon!

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Marijn "i before j" van Putten


Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @PhDniX

28 Sep
Great conversation between @dbru1 and Asma Hilali about quranic manuscripts but to me one detail remained a bit vague, it is addressed in the title: "Did the Quran exist early as a book?"

The answer to this should, unequivocally be: Yes. Yes it did.

Thread 🧵
One of the questions posed in the conversation is "where is Uthman's codex?" and "where are the regional codices?"

We might actually have them, but the fragments we have simply do not come with labels.

But even if we didn't this does not mean they aren't CERTAINLY a reality.
We do not own the autograph of Sībawayh's al-Kitāb, are we to assume al-Kitāb never existed and Sībawayh did not write it? Of course not. And this is true for the vast majority of the Classical Arabic literature (or any literature in manuscript traditions).
Read 16 tweets
24 Sep
It's been a while since I wrote on #Tamazight/#Berber, so I thought it would be nice to do a thread on its plural formation. Much like Arabic (and Semitic more generally), Berber has 'sound plurals' and 'broken plurals'. Let's have a look at how these are reconstructed.
The 'sound plural' is simply formed by a suffix: *-ăn for the masculine, *-en for the feminine. For example:
*a-maziɣ pl. i-maziɣ-ăn "Berber man"
*ta-maziɣ-t pl. *ti-maziɣ-en "Berber woman"

In most dialect *-ăn > -ən and *-en > -in, but some retain the contrast.
If a sound plural suffix is added to a word-final vowel, usually an epenthetic *-t- is infixed to avoid the meeting of two vowels. Perhaps this is due to historical loss of *t (there is some weird stuff with disappearing *t's elsewhere in morphology).

*anu pl. *anu-t-ăn 'well'
Read 19 tweets
9 Sep
The Arab grammarians describe the occurrence of /ē/ (called ʾimālah) in word-final position of verbs and nouns, e.g. banē "he built" and ḥublē "pregnant". This thread discusses the disagreements between different groups of grammarians, and how the opinions develop. 🧵
The earliest and most famous grammarian Sībawayh (d. 180 AH) tells us: nouns which in Classical Arabic end in -ā, whose third root consonant is yāʾ undergo ʾimālah, while those whose third root consonant is wāw do not, thus: al-hudē "the guidance" but al-ʿaṣā "the stick".
One is sure to notice that this distribution aligns with how the final -ā is spelled. If it is spelled with yāʾ it is pronounced -ē, if it is spelled with ʾalif it is spelled -ā. This is certainly not a coincidence, the spelling seems to retain a memory of this distinction.
Read 14 tweets
9 Sep
In case anyone who cares about my bad fighting game takes among my followers, give me some likes! Image
1. Probably The King of Fighters 2002UM, though due to bad netplay I haven't played it nearly as much as OG2k2. After that probably Alpha 2.
2. Hard to say. Got burned out the most on Street Fighter 4. But still enjoyed watching it after. Alpha 3 is incredibly ugly and terrible. Have a love-hate relation with SF3: Third Strike.
Read 26 tweets
3 Sep
First time that the pre-Islamic record gives evidence of fully functioning Tanwīn! Even though the use of full ʾiʿrāb + tanwīn becomes THE prime feature for high "Classical" speech in Arabic, up until we would only find it in Islamic Era Classical Arabic.
Forms of Old Arabic in the pre-Islamic period consistently lack this feature. It is absent Safaitic, Hismaic and Nabataean Arabic.
Some people, like Jonathan Owens, even doubted that it was a genuine feature at all, and it was maybe made up by the Arab Grammarians.
Based on historical linguistic evidence, there was reason to believe the feature was archaic, even though it only showed up in late sources, as me and @Safaitic have argued in our reply to Jonathan Owens here:
Still, it is good to now have evidence for it!
Read 9 tweets
1 Sep
When the 3rd caliph ʿUṯmān b. ʿAffān tasked the committee led by Zayd b. Ṯābit to write the official Mushaf, he is said to have told them "write it in the language of the Quraysh, for it has been revealed in their tongue." This seems incompatible with the canonical readings. 🧵
While today there are 10 accepted canonical readings, with 2 transmissions each which can differ quite significantly in their linguistic details, not a single one of them agrees with the universally agreed upon linguistic features of Hijazi (or more specifically Qurashi) Arabic.
There is consensus on the fact that the Quraysh did not have vowel harmony of the third person masculine suffixes -hū and -hum.
Farrāʾ (d. 209) reports Quraỵš/Hijazis say: ʿalayhum, ʿalayhumā, ʿalayhunna, ʿalayhu, fīhu, bihū, while Najdis say ʿalayhim, -himā, -hinna, -hi, -hī.
Read 12 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!