Thread by @PhDniX

A recurring criticism on the work on the history of Arabic done by @Safaitic and me is that we rely too much on Classical Arabic, and ignore the modern dialects.

We both have done historical works on the dialects, and they are essential for the history of Proto-Arabic. A thread:

1) Barth-Ginsberg alternation: In Classical Arabic the prefix vowel of the verb is always a, e.g. yaktubu, yaqtilu, yaʕrafu. From a comparative perspective this is surprising. For Hebrew and Aramaic it seems that before *a the prefix vowel was i, i.e. *yiʕrafu.

Bloch already showed brilliantly that there are clear signs of this alternation in the descriptions by the medieval grammarians for Old Arabic, and connects it to Hebrew. Therefore there is already no doubt that Proto-Arabic had *yaktubu but *yiʕrafu. menadoc.bibliothek.uni-halle.de/dmg/periodical…

The fact that it aligns perfectly with the Hebrew/Aramaic system makes it clear that the grammarians weren't making it up. But it can be confirmed independently through Arabic dialects. In Najdi Arabic the system is preserved, as described by Bruce Ingham. books.google.nl/books?id=t1W7s…

In a forthcoming article I argue that traces of this same alternation can also be found in Maltese and Tunis Arabic. Classical Arabic, in this regard, is clearly innovative, having lost the alternation. Classical Arabic cannot be ancestral to these dialects.

Traces of this are also found in Old Arabic of the epigraphic record which I am sure @safaitic will expand upon (I also thank him for alterintg me of the presence of the situation in Najdi Arabic).

2) Retention of an Old Arabic fourth vowel: Grammarians distinguish two different vowels an ā and an ē, which are clearly distinct in Quranic Arabic, as can be deduced from the rhyme. Those with /ā/ are written with alif: daʕā دعا, those with /ē/ are written with yāʾ: ramē رمى

I have written extensively on this in my article on the Proto-Arabic triphthongs. The ā comes either from *ā or *awa; while the ē comes from *aya. In Safaitic these are still distinct triphthongs, and continue a Proto-Semitic triphthong academia.edu/32715681/The_d…

All trace of this is lost in Classical Arabic as well as in most modern Arabic dialects. But as Behnstedt (1987: 133-5) shows that the Yemeni dialect known as ar-Rāziḥīt retains an ā vs. ē distinction, an archaism compared to Classical Arabic. books.google.nl/books?id=pvKmQ…

3) The shape of the prefix of Stem V and VI verbs: In Classical Arabic the prefix of these stems is always ta-, i.e. takallama, yatakallamu. In most modern dialects the prefix is simply t-, i.e. itkallam, yitkallam. While some dialects retain a reflex of *ta-.

As already clearly observed by Diem the itkallam forms cannot be explained as coming from the Classical Arabic form through regular sound development. He suggests that the Ethiopic pattern: täḳättälä, yətḳättäl is original and that CAr. and Dialects generalized different forms.

This is a convincing argument, and disassembled the tired Fergusonian "whenever it is not Classical Arabic, it must mean the dialects did something wrong". You can read the full article here: menadoc.bibliothek.uni-halle.de/dmg/periodical…

What Diem did not see is that the several modern dialects retain exactly this alternation that he predicted for Proto-Arabic on good comparative semitic grounds. It is found in several Bedouin-type dialects, e.g. Douz Arabic (southern Tunisia), Gulf Arabic and Dōsiri.

4) The diptotic feminine ending: In Classical Arabic, the feminine ending *-at- is triptotic, it takes nunation and three cases in the indefinite. This seems to have also been the case for Quranic Arabic and presumably Classical Arabic. See my article: academia.edu/35131582/_The_…

While there are plenty of ways that people have tried to explain Quranic (and Classical Arabic's) odd orthography and behaviour of this ending, dialectal Arabic seals the deal, and retaining the Proto-Arabic diptotic system it seems to point to.

Tihami Arabic retains a distinction between original triptotes (ending in -u or un) and diptotes (no ending). In Tihami Arabic, the feminine ending never takes the ending -u(n), confirming that it was not a triptote but a diptote, much like Quranic Arabic and probably Nabataean.

This, once again, is irreconcilably different from Classical Arabic, and in this case Classical Arabic is clearly the innovative variant while Quranic and Tihami Arabic are archaic.

Shammari Arabic also has a very archaic treatment of the feminine ending: academia.edu/18687778/_The_…

5) The etymology of ḥattā 'until': @Safaitic has shown that a crucial element of understanding the etymology of ḥattā rests solely on the Levantine dialects, and could not have been retreived from Classical Arabic data. academia.edu/35609082/Al-Ja…

6) The etymology of Levantine pronouns in -e: @Safaitic argues that the Levantine pronouns that end in -e provide crucial evidence for an ancient Semitic pronominal *-h suffix. Little (if any) evidence of this is to be found in Classical Arabic. academia.edu/7542944/Al-Jal…

7) Voiceless reflexes of ظ and ض in Pre-Hilali Maghrebi Arabic: @Safaitic has shown that in Old Arabic the ظ and ض were voiceless ejectives. He finds traces of evidence for this in Maghrebi Arabic. academia.edu/20913515/Al-Ja…

One might add that in Im-Maṯ̣ṯ̣ah and Rāziḥit (both Yemen) voiceless reflexes of these phonemes are also attested.

8) Lateral realization of ش and ض: The reason why Arabic is famously called luġat al-ḍād "The language of the ḍād", is because in Classical Arabic it appears to have originally had a very exotic lateral fricative pronunciation. This is retained in good recitation.

Etymologically both the ش and ض were laterals. @Safaitic has argued that in Old Arabic (besides the ض) also ش was still a lateral. This is proven beyond doubt by Rāziḥīt simply retains a lateral realization for both as Watson shows academia.edu/7583140/Al-Jal… jstor.org/stable/4122387…

9) Evidence for an affricate ص: @Safaitic has suggested that Sībawayh describes an affricated tṣ pronounciation for the ص in Classical Arabic. Traces of this appear to present in im-Maṯ̣ṯ̣ah's realization of the ص as [st], e.g. mast < *maṣṣa

academia.edu/8770050/Al-Jal…

To conclude, the inclusion of dialectal data is essential for understanding the history of Arabic. Old-school Arabists assumed the dialects came from Classical Arabic. I have just shown that this is not the case. This view is unduly attributed to "The New School of Old Arabic".