Arabic Grammar – 13

►II.21.A.e. THE GUTTURALS x, ɣ, ħ, ʕ

§α. The pharyngeal fricatives ħ ح and ʕ ع are typical of the Afro-Asiatic languages and they are thought to have existed in the Proto-Afro-Asiatic language. However, these sounds were lost historically from Akkadian, Ethiopian languages, and Coptic, and they were lost prehistorically from the Berber languages (and the Indo-European languages). In most of these cases the loss of the pharyngeals seems to have been a substrate influence (i.e. it was caused by the imposition of the language on many non-native speakers whose languages lacked similar sounds).

The loss of the pharyngeals is rare in Arabic, but it is attested in some dialects of southern Yemen (e.g. in the speech of Bedouin women of the hamlet of Θuwa in ʔAbyan, see Martine Vanhove, Yemen in Encyclopedia of Arabic Language and Linguistics).

§β. It is interesting that Sibawayh places the point of articulation of x خ and ɣ غ further back in the vocal tract than the point of articulation of q ق , which in turn is further back than the point of articulation of k ك (vide §II.20.B.). In chapter 570 Sibawayh says the following:

وَالْخَاءُ والْغَيْنُ بِمَنْزِلَةِ الْقَافِ، وَهُمَا مِنْ حُرُوْفِ الْحَلْقِ بِمَنْزِلَةِ الْقَافِ مِنْ حُرُوْفِ الْفَمِ وَقُرْبُهُمَا مِنْ الْفَمِ كَقُرْبِ الْقَافِ مَنْ الْحَلْقِ.

Xȃʔ and ɣayn are analogous to qȃf, they are to the throat letters like the qȃf is to the mouth letters, their closeness to the mouth is like the closeness of qȃf to the throat.

If we accept what Sibawayh says, we should presume that xȃʔ and ɣayn were uvular fricatives, qȃf was a uvular plosive, and kȃf was a velar plosive.

In almost all the Semitic languages other than Arabic, the sounds which correspond to Arabic x خ and ɣ غ merged with the pharyngeal ħ ح and ʕ ع . This suggests that the original Semitic sounds were uvular χ and ʁ. If Sibawayh’s description is accurate, it seems that the original uvular quality of the Arabic sounds was still preserved in the 8th century AD, at least in some dialects.

§γ. According to Ahmad Al-Jallad (An Outline of the Grammar of the Safaitic Inscription, (2015) p. 43, The Petra Papyri II. Amman: American Center of Oriental Research, pp. 23-48, and Graeco-Arabica I: The Southern Levant (2015)) the sounds x, ɣ, ħ, ʕ, h and ʔ are usually not written in the Greek transcriptions of Ṣafaitic words and in the pre-Islamic papyri. Thus, the spelling Αλιζου (Alizou) stands for Ṣafaitic Xlṣ, Ανεμος (Anemos) stands for Ṣafaitic Γnm, and νααρ (naar), αρβαθ (arbath), and Αλεβους (Alebous) in the papyri stand for *nahar “torrent, gully,” *xarbat “ruin(s),” and the name Γȃlibuṋ غَاْلِبٌ .

It seems that the sounds x and ɣ were judged close to the laryngeal h, and so they were left unwritten, because this sound is not written in the Greek orthography (it is indicated by a diacritic called the spiritus asper). That the sounds x and ɣ were distinguished from the pharyngeal ħ and ʕ is proved by the presence of some spellings with Greek χ and γ. These are, according to Al-Jallad, attested in the epigraphy and papyri from all regions and time periods, e.g. Χαλιπος (Chalipos) for Ṣafaitic Xlf and Γαυθος (Ghauthos) for Ṣafaitic Γθ.

§δ. It is interesting that the Arabic ħ ح and ʕ ع sometimes correspond to Ancient Egyptian x , which is thought by some to have sounded χ (Antonio Loprieno (1995), Ancient Egyptian: A linguistic introduction). For example, Arabic raħaḍa رَحَضَ “washed” corresponds to Ancient Egyptian rt “wash clothes,” Arabic malȋħuṋ مَلِيْحٌ “handsome” (in Aleppo mnȋħ “good”) corresponds to Ancient Egyptian mn “excellent; effective,” Arabic wasiʕa وَسِعَ “be(came) wide” corresponds to Ancient Egyptian ws “be(come) wide,” and Arabic sabʕuṋ سَبْعٌ “seven” corresponds to Ancient Egyptian sf “seven.” Apparently, Ancient Egyptian lacked a voiced counterpart of x (some scholars maintain that x was voiced and that X was its voiceless counterpart, however, it seems that both of these were voiceless and X was a fronted variant of x). It appears that the voiced *ʁ merged in pre-Ancient Egyptian with the voiceless x , hence, it is not surprising if this sound corresponds to both of Arabic ħ ح and ʕ ع . Apart from voice, scholars are in dispute over explaining the discrepancy in the point of articulation between the Arabic and Egyptian forms. One point of view considers the Arabic forms more conservative, even when there is internal evidence within Semitic to the contrary. Thus, the ħ in Arabic rħ “wash” is considered more original than the χ in both Akkadian rχ “wash” and Ancient Egyptian rt.

In fact, there are many Akkadian roots which have χ corresponding to Arabic ħ, e.g. Akk. χpr vs. Ara. ħfr “dig,” Akk. nbχ vs. Ara. nbħ “bark,” Akk. χ vs. Ara. msħ “measure,” Akk. χyṭ vs. Ara. ħyṭ “guard,” Akk. pχl “breed an animal” vs. Ara. fħl “stallion,” Akk. nχt vs. Ara. nħt “trim, clip,” Akk. χbb “caress” vs. Ara. ħbb “love” (dialectal “kiss”), etc.. Sometimes Akkadian has a form with ħ as well, e.g. Akk. χbl and *ħbl vs. Ara. ħbl “rope,” Akkadian lχw and *lħy vs. Ara. lħy “jaw, cheek,” etc..

In my opinion, the simplest explanation for all these discrepancies might be the well-attested Semitic tendency for the changes χ > ħ and ʁ > ʕ. In several West Semitic languages the uvular *χ and *ʁ merged completely with the pharyngeal ħ and ʕ. This happened in Hebrew and Aramaic in the late 1st millennium BC, in Ethiopian languages, and in Modern South Arabian languages. It is not impossible that the same tendency existed (partly) in early Arabic, and also in pre-Proto-Semitic.

It seems that the older languages tend to have more *χ and *ʁ sounds than the younger ones, that is to say, Akkadian has more of these sounds than Arabic, and Ancient Egyptian has more of them than Akkadian.

After the sounds and were fronted in Arabic to the velar x and ɣ, they became stable and the merger with the pharyngeals subsided.

An alternative explanation was suggested by John Huehnergard. He suggested that there was in Proto-Semitic an emphatic velar fricative *x̣. This became x in Akkadian but ħ in West Semitic.



Comparison with Berber and Cushitic languages cannot prove that Semitic forms are more conservative than Ancient Egyptian ones, because it seems likely that the Berber and Cushitic languages are more closely related to Semitic than to Ancient Egyptian. It seems that the Semitic, Berber, and Cushitic languages are descended from a single genetic node in the Afro-Asiatic tree. Thus, similarity between these three groups cannot be taken as evidence against the conservativeness of Ancient Egyptian forms. Besides, the Berber and Cushitic languages are (practically) only attested in recent times, whereas Ancient Egyptian is the oldest attested Afro-Asiatic language.


►II.21.A.f. THE EMPHATICS ð̣, ṭ, ṣ, ḍ, q

§α. The description of the muṭbaq (emphatic) letters by Sibawayh (§II.20.J.) does not say a word about constricting the throat. In fact, the pharyngealization of the emphatic consonants is uncertain even today. Some authors have described the Arabic emphatic consonants as uvularized rather than pharyngealized. Some sources claim that the emphatic consonants in Arabia and Mesopotamia are merely velarized.

The phoneme which corresponds to Arabic q ق in Ethiopic and the Modern South Arabian languages is an ejective velar stop , and there are indications that the same pronunciation existed also in other Semitic languages. It is believed that q ق was originally the emphatic counterpart of k ك and g ج . The fact that * changed in Arabic to a uvular stop q perhaps indicates that it was uvularized *. It makes sense for a uvularized consonant to change to a uvular consonant. A similar uvularization of a uvularized consonant is perhaps behind the Aramaic change *ṣ₂ > <q> > ʕ (vide infra §II.21.A.f.π.).

§β. One of the bad supplementary letters mentioned by Sibawayh (§II.20.A.c.) is the “ṣȃd which is like sȋn” اَلْصَّاْدُ اَلَّتِيْ كَاَلْسِّيْنِ . This perhaps refers to a weakly emphasized variant of Sibawayh’s correct ṣȃd. If Sibawayh’s correct ṣȃd was uvularized or pharyngealized , the weaker variant could have been velarized .

According to Sibawayh (chapter 569 ), sȋn was changed to ṣȃd when it was followed in the same word by q. For example, ṣuqtu صُقْتُ for suqtu سُقْتُ “(I) drove” and ṣabaqtu صَبَقْتُ for sabaqtu سَبَقْتُ “(I) outran.” In modern Aleppo Arabic (and other dialects) there appears to be an opposite sound law to that which Sibawayh stated. In words containing q there is a dissimilated s < ṣ. For example, saddaʔ̣ for ṣaddaqa صَدَّقَ “(he) believed,” safaʔ̣ for ṣafaqa صَفَقَ “(he) slapped,” and sʔ̣ȋʕ for ṣaqȋʕa صَقِيْعٌ “freeze.”

Could it be that the “ṣȃd which is like sȋn” which Sibawayh mentioned as one of the bad supplementary Arabic letters is this dissimilated s < ṣ? In my opinion, this is not very likely, because the dissimilated s < ṣ sounds just like the regular s, so it is not an “additional letter.”

§γ. There is much indirect evidence that the Akkadian emphatic consonants were ejective consonants. Such a pronunciation of the emphatics exists in the Ethiopian Semitic and Cushitic languages, and more importantly in the Modern South Arabian languages (where a Cushitic substrate influence is usually ruled out). It is widely believed that the Proto-Semitic emphatics were ejective and that this pronunciation changed in Arabic (and likely also in the Northwest Semitic languages) to uvularization or pharyngealization.

§δ. Ejective consonants can only be voiceless. It is believed that the Semitic emphatics were originally voiceless. However, in Arabic the emphatics are commonly voiced.

According to Sibawayh (§II.20.C.), all the emphatic consonants save ṣȃd are maǵhȗr (voiced). Sibawayh says that without the ʔiṭbȃquṋ إِطْبَاْقٌ “closing” (emphasis) ṭȃʔ will become dȃl, ṣȃd will become sȋn, ð̣ȃʔ will become ðȃl, and ḍȃd will vanish because there is no other sound that shares its [primary] point of articulation.

وَلَوْلَا الْإِطْبَاقُ لَصَارَتْ الْطَّاءُ دَالًا، وَالْصَّادُ سِينًا، وَالْظَّاءُ ذَالًا، وَلَخَرَجَتْ الْضَّادُ مِنْ الْكَلَامِ لِأَنَّهُ لَيْسَ شَيْءٌ مِنْ مَوْضِعِهَا غَيْرُهَا.

§ε. Sibawayh repeats over and over that ṭȃʔ ط is maǵhȗr, and that it is the muṭbaq (emphatic) counterpart of dȃl د .

كَمَا أَنَّ الْطَّاءَ لَيْسَ حَرْفٌ أَقْرَبَ إِلَيْهَا وَلَا أَشْبَهَ بِهَا مِنْ الْدَّالِ

[…] like how there is no letter closest to ṭȃʔ and more similar to it than dȃl (chapter 567 )

إلا أنَّ إذهابَ الإطباقَ [من الطاءِ عند إدغامها] مع الدالِ أمثلُ قليلًا [من إذهابه عند إدغامها مع التاءِ]، لأنَّ الدالَ كالطاءِ في الجَهْرِ، والتاءُ مهموسةٌ.

[…] but removing ʔiṭbȃq [from ṭȃʔ when it is assimilated] to dȃl is a bit more ideal [than removing it from ṭȃʔ when it is assimilated to tȃʔ], for dȃl is maǵhȗr like ṭȃʔ, [but] tȃʔ is mahmȗs (chapter 568 )

[والصادُ] من السينِ كالطاءِ من الدالِ، لأنها مهموسةٌ مثلُها وليس يفرق بينهما إلا الإطباقُ، وهي من الزايِ كالطاءِ من التاءِ، لأن الزايَ غيرُ مهموسةٍ

[ṣȃd] to sȋn is like ṭȃʔ to dȃl, for it is mahmȗs like it and there is no difference between them save ʔiṭbȃq, and [ṣȃd] to zȃy is like ṭȃʔ to tȃʔ, for zȃy is not mahmȗs (vchapter 568 )

The linguists who followed Sibawayh unanimously adopted his classification of ṭȃʔ as maǵhȗr. ʔIbn Yaʕȋš اِبْنُ يَعِيْشَ (died in 1245 AD) says in his commentary on مُفَصَّلُ الْزَّمَخْشَرِيِّ (in فصل حروف العربية ) that the ṣȃd in the Qurʔȃnic word ṣirȃṭuṋ صِرَاْطٌ “path” was pronounced by some Qurʔȃn readers , and he explains this as a partial assimilation of the ṣȃd to the ṭȃʔ, because both sounds are maǵhȗr.

وكذلك الصاد التي كالزاي، نحو قولهم في “مصدر”: “مصدر”، وفي “يصدق”: “يصدق”، وقد قُرىء {الصِّرَاطَ الْمُسْتَقِيمَ} بإشمام الصادَ الزايَ، وهي قراءة حمزة. وعن أبي عمرو فيها أربع قراءات، منها “الصراط” بين الصاد والزاي، رواها عُرْيان بن أبي شَيْبانَ، قال: سمعتُ أبا عمرو يقرأ: “الصراط” بين الصاد والزاي، كأنّه أشرب الصادَ صوتَ الزاي حتى تُوافِق الطاء في الجهر، لأن الصاد مهموسة، والطاء والدال مجهورتان، فبينهنّ تنافٍ وتنافرٌ، فأشربوا الصاد صوت الزاي, لأنّها أختُها في الصفير والمخرج، وموافقةٌ للطاء والدال في الجهر، فيتقارب الصوتان، ولا يختلفان. [شرح المفصل لابن يعيش. فصل حروف العربية]

However, in the same source we find the following:

وأمّا الطاء التي كالتاء، فإِنَّها تُسْمَع من عَجَم أهل العراق كثيرًا، نحو قولهم في “طالب”: “تالب”؛ لأنّ الطاء ليست من لغتهم، فإذا احتاجوا إلى النطق بشيء من العربيّة فيه طاءٌ، تكلّفوا ما ليس في لغتهم، فضعفُ لفظُهم بها. [شرح المفصل لابن يعيش. فصل حروف العربية]

As for the ṭȃʔ which is like tȃʔ, this is commonly heard from the non-Arabs in Iraq, like when they say tȃlibuṋ تالب for ṭȃlibuṋ طالب [“student”]. Since ṭȃʔ is not part of their language, if they need to say something Arabic which has ṭȃʔ, they will be adopting something which is not in their language, and so they will pronounce it weakly.

If ṭȃʔ was pronounced in Iraq, those non-Arabic-speakers should have pronounced it d. It is strange that the author does not say anything about this discrepancy.

The linguist ʔIbn ʔ·al-Ǵazariyy اِبْنُ الْجَزَرِيِّ who lived in the 14-15th centuries says (in التمهيد ) the following while talking about the pronunciation of ḍȃd:

واعلم أن هذا الحرف ليس من الحروف حرف يعسر على اللسان غيره، والناس يتفاضلون في النطق به. فمنهم من يجعله ظاء مطلقاً […] وهم أكثر الشاميين وبعض أهل المشرق […] ومنهم من لا يوصلها إلى مخرجها بل يخرجها دونه ممزوجة بالطاء المهملة، لا يقدرون على غير ذلك، وهم أكثر المصريين وبعض أهل المغرب. [ابن الجزري، التمهيد في علم التجويد، فصل نذكر فيه ما يتعلق بكل حرف من التجويد]

Know that this is the only letter that is difficult to pronounce, and people pronounce it with varying success. Some of them make it always a ð̣ȃʔ […] those are most of the [Greater] Syrians and some of the people of the Levant […] Others do not deliver it to its point of articulation but they let it out from a nearer point, mixed with ṭȃʔ. They cannot do better than that. Those are most of the Egyptians and some of the people of the Maghreb.

The pronunciation of ḍȃd in Egypt which the author means must be the same as the modern pronunciation there, which is . It is significant that he does not call this sound ṭȃʔ, but rather he describes it as a sound “mixed with ṭȃʔ.”

I am not convinced that Sibawayh and those who followed him were describing a really voiced ṭȃʔ. As far as I know, a completely voiced ṭȃʔ has never been attested in any Arabic dialect, nor a voiced ṭȃʔ has been attested in foreign transcriptions of Arabic words or vice versa.

Allophonic voicing of ṭȃʔ (as well as tȃʔ) in pre- and inter-vocalic positions is known in central Yemen, and it is found sporadically in other dialects (e.g. ndaq for naṭaqa نَطَقَ “(he) spoke” in the dialect of Cherchell).

Given that Sibawayh was identifying the maǵhȗr sounds based on the breath that accompanied them, it is not surprising that he would classify the voiceless ṭȃʔ as a maǵhȗr sound, because the voiceless ṭȃʔ, like the voiced dȃl, is unaspirated, whereas the voiceless tȃʔ is aspirated. Because ṭȃʔ did not have a voiced counterpart with which Sibawayh could compare it, he could easily have made a mistake and connected it with the similarly breathless dȃl.

A similar mistake is still made nowadays. People have claimed that a voiced ṭȃʔ exists in southern Egypt, but the real sound that exists there is apparently a glottalized ˀ (see Egypt in the Encyclopedia of Arabic Language and Linguistics). A weakly glottalized sound can be confused for a voiced one, because it is unaspirated. Glottalized emphatics of the Modern South Arabian languages have been frequently called voiced. Since the glottalization of the Modern South Arabian emphatics is said to be very weak, this could be the reason.

§ζ. In the Syriac language the letters ܛ and q ܩ were regularly used for writing Greek τ ·t and κ (k). A similar usage is found in the Ṣafaitic inscriptions and the pre-Islamic papyri, e.g. Ṣafaitic Mrṭs₁ for Μύρτος (Mýrtos) and ʔqlds₁ for Claudius. In Ṣafaitic the letter is frequently used for transcribing Greek σ and Latin s, e.g. Ṣafaitic Flfṣ for Φίλιππος (Phílippos), Grfṣ for Ἀγρίππας (Agríppas), Hrdṣ for Ἡρῴδης (Hērōídēs), and Grmnqṣ for Germanicus. A similar usage is found in the Sabaic Qṭwṣf for Κτησιφῶν (Ktēsiphȏn). The reason why the emphatics ṭ, ṣ and q were used for transcribing Greek τ, σ and κ is probably because the Greek sounds were unaspirated, hence they sounded close to the Semitic emphatics, which were voiceless and unaspirated. This also explains why Greek θ (th) and χ (ch) were used in Ṣafaitic and the pre-Islamic papyri for writing Arabic t and k. Since Greek θ and χ originally represented the aspirated stopsand , they were felt closer to the Arabic t and k than the plain voiceless τ and κ.

The same usage was common with Iranian words in early Classical Arabic, e.g. Ṭabaristȃnu طَبَرِسْتَاْنُ , Ṭuxaristȃnu طُخَاْرِسْتَاْنُ , Ṭȗsu طُوْسُ , ʔIṣṭaxru إِصْطَخْرُ , ʔIṣfahȃnu إِصْفَهَاْنُ , Qumu قُمُ , Qazwȋnu قَزْوِيْنُ , etc.

The use of ط , ṣ ص and q ق for transcribing foreign t, s and k became a tradition in Classical Arabic. It was used always, regardless of the actual pronunciation. The modern authorities of Arabic apparently still prescribe this usage, but since it is not taught at schools, many modern Arabic-speakers are unaware of it. Hence, foreign t, s and k are nowadays commonly transcribed with the plain t ت , s س and k ك .

§η. One of the bad supplementary letters mentioned by Sibawayh (§II.20.A.c.) is the “ṭȃʔ which is like tȃʔ” اَلْطَّاْءُ اَلَّتِيْ كَاَلْتَّاْءِ . This could be a reference to the voiceless if we suppose that Sibawayh’s correct ṭȃʔ was voiced . However, it would be very strange if Sibawayh considered the extremely common sound a bad letter not used in reciting Qurʔȃn or poetry. More likely, the “ṭȃʔ which is like tȃʔ” means a weakly emphasized , perhaps a velarized tˠ if we assume that Sibawayh’s correct ṭȃʔ was uvularized or pharyngealized .

In the modern dialects is de-emphasized to t in some words, but this is probably not the supplementary letter which Sibawayh meant, because it sounds just like the regular t.

§θ. Another one of the bad supplementary letters (§II.20.A.c.) is the “weak ḍȃd” اَلْضَّاْدُ اَلْضَّعِيْفَةُ . It is unclear what Sibawayh meant by this. His description of it is of no great help:

إلا أن الضاد الضعيفة تتكلف من الجانب الأيمن، وإن شئت تكلفتها من الجانب الأيسر وهو أخف، لأنها من حافة اللسان مطبقة، لأنك جمعت في الضاد تكلف الإطباق مع إزالته عن موضعه‏.‏ وإنما جاز هذا فيها لأنك تحولها من اليسار إلى الموضع الذي في اليمين‏.‏ وهي أخف لأنها من حافة اللسان وأنها تخالط مخرج غيرها بعد خروجها فتستطيل حين تخالط حروف اللسان، فسهل تحويلها إلى الأيسر لأنها تصير في حافة اللسان في الأيسر إلى مثل ما كانت في الأيمن، ثم تنسل من الأيسر حتى تتصل بحروف اللسان كما كانت كذلك في الأيمن‏.

From this description it appears that the “weak ḍȃd” is also a lateral consonant like the correct ḍȃd. It can be articulated either on the right side or the left side of the tongue. Perhaps this was just a weakly emphasized variant of the correct ḍȃd (i.e. a velarized ɮˠ). Since Sibawayh did not know an unemphatic counterpart of ḍȃd, he could not describe this variant with a phrase like “the ḍȃd which is like…” Hence, he just said “the weak ḍȃd.”

Following is what ʔIbn Yaʕȋš has on Sibawayh’s weak ḍȃd. What he says is probably incorrect, because the weak ḍȃd described by Sibawayh is obviously a lateral consonant like the correct ḍȃd.

والضاد الضعيفة من لغةِ قوم اعتاصت عليهم، فرُبما أخرجوها ظاء، وذلك أنّهم يُخْرِجونها من طرف اللسان وأطرافِ الثنايا، وربّما راموا إخراجَها من مخرجها، فلم يَتَأَتَّ لهم، فخرجت بين الضاد والظاء.

The weak ḍȃd belongs to the language of people who have great difficulty with it, and so they may pronounce it ð̣ȃʔ, that is, they let it out from [between] the tip of the tongue and the tips of the front teeth. They may want to let it out from its correct point of articulation, but they cannot, and so it becomes between ḍȃd and ð̣ȃʔ.

§ι. Another one of the bad supplementary letters (§II.20.A.c.) is the “ð̣ȃʔ which is like θȃʔ” اَلْظَّاْءُ اَلَّتِيْ كَاَلْثَّاْءِ . This is apparently a reference to a voiceless θ̣, but I do not know if this was a phoneme or just an allophone. A voiceless θ̣ phoneme is perhaps attested in Ancient North Arabian languages (vide §II.21.A.b.β.).

ʔIbn Yaʕȋš had nothing to say about Sibawayh’s “ð̣ȃʔ which is like θȃʔ.” He just gave an example (which is perhaps made up by him?)

ومثالُ الظاء كالثاء قولهم في “ظلم” “ثلم”.

An example of the ð̣ȃʔ which is like θȃʔ is θalama ثَلَمَ for ð̣alama ظَلَمَ [“(he) was unjust”].

§κ. One of the supplementary Arabic letters which Sibawayh considered good (§II.20.A.b.) is the “ṣȃd which is like zȃy” اَلْصَّاْدُ اَلَّتِيْ تَكُوْنُ كَاَلْزَّاْيِ . The explanation of this sound is perhaps found chapter 569 of Sibawayh’s book. Here is described a voiced allophone of ṣȃd that appeared in words such as maẓdaruṋ for maṣdaruṋ مَصْدَرٌ “source,” ʔaẓdartu for ʔaṣdartu أَصْدَرْتُ “(I) sent out,” and taẓdȋruṋ for taṣdȋruṋ تَصْدِيْرٌ “exporting.” Some speakers (whose speech was highly esteemed by Sibawayh) even said tazdȋruṋ and ʔazdartu, with a plain z.

وَسَمِعْنَا الْعَرَبَ الْفُصَحَاءَ يجعلونها زايًا خالصةً، كما جعلوا الإطباقَ ذاهبًا في الإدغام‏،‏ وذلك قولكَ في التَّصْدِيرِ‏ “التَّزْدِيرُ”، وفي الفَصْد‏‏ِ “الفَزْدُ”، وفي أَصْدَرْتُ‏ “أَزْدَرْتُ”‏.‏

Besides the allophone which Sibawayh described, it is possible that some speakers always pronounced ṣȃd with voice. This would explain why some Qurʔȃn readers said ẓirȃṭuṋ for ṣirȃṭuṋ صِرَاْطٌ “path” (vide supra §II.21.A.f.ε.), and it would (possibly) explain why Ṣafaitic was (apparently) rendered by Greek ζ (z) in []Αλιζου ([H]alizou) for Ṣafaitic Xlṣ, and possibly σειαζ (seiaz) for Ṣafaitic s₂ħṣ (see Ahmad Al-Jallad (2015), An Outline of the Grammar of the Safaitic Inscription, p. 43).

§λ. Sibawayh characterized qȃf as a voiced uvular plosive, i.e. ɢ. This pronunciation is commonly found today in Yemen, where it often alternates with q (as an allophone or a free variant). It is also found (non-dominantly) in eastern Arabia (the Gulf) and in Sudan (e.g. Sudanese ɢānūn for qȃnȗnuṋ قَاْنُوْنٌ “law” and laɢab for Classical laqabuṋ لَقَبٌ “cognomen”). This is also the Persian pronunciation of Arabic ق .

§μ. In all the dialects where the pronunciation ɢ for ق exists, there are claims that the pronunciation ɣ for ق also exists. The sound ɢ is a relatively rare sound in world languages and people can mishear it as a voiced uvular fricative ʁ. However, it is also likely that ق is actually pronounced ʁ or ɣ in some dialects (at least in some words). A phonetic change (*ɣ >) *ʁ > *ɢ > q is known in some rural and Bedouin dialects of northwestern Iraq, where غ is pronounced q, while ق is pronounced g, thus they say qanam for ɣanamuṋ غَنَمٌ “sheep” but gām for qȃma قَاْمَ “(he) rose, stood up” (Otto O. Jastrow, Iraq in Encyclopedia of Arabic Language and Linguistics). The same change is also found sporadically in the Gulf Arabic, e.g. qēr for ɣayruṋ غَيْرٌ “other” and qanna for ɣannȃ غَنَّىْ “(he) sang.”

§ν. In the Bedouin dialects ق is typically pronounced g. The pronunciation g for ق is nowadays dominant in Saudi Arabia, the Gulf, in the Bedouin dialects of the northern Bȃṭina·t of Oman, and in Ħaḍramawt. This pronunciation is also typical of the gələt dialects of Mesopotamia, and is commonly found in Egypt, Sudan and the Maghreb (wherever it is found, it is due to Bedouinic influence).

§ξ. The origin of the voiceless pronunciation q appears to be eastern Arabia (see Chaim Rabin (1951), “Ancient West-Arabian,” p. 125). The following is from الصاحبي by the Iranian linguist ʔIbn Fȃris اِبْنُ فَاْرِس (died in 1004 AD):

فَأَمَّا بَنُو تَمِيْمٍ فَإِنَّهُمْ يُلْحِقُونَ الْقَافَ بِاَلْلَّهَاةِ حَتَّى تَغْلُظَ جِدَّا، فَيَقُولُونَ “الْقَوْمُ” فَيَكُونُ بَيْنَ الْكَافِ وَالْقَافِ، وَهَذِهِ لُغَةُ فِيهِمْ.

As for Tamȋm, they push the qȃf towards the uvula so that it becomes very rough. They say ʔ·al-qawmu الْقَوْمُ [“the people”] so that [the qȃf ] is between kȃf and qȃf. This is a language among them.

In modern Arabia the pronunciation q for ق is found in Yemen, where it alternates with ɢ/ʁ (sometimes as an allophone, and sometimes as a free variant) and in Oman (in the capital area, most of the Bȃṭina·t coast, the bigger towns on the mountain fringes, and some mountain villages). This pronunciation is commonly found in the qəltu dialects of sedentary northern Mesopotamia, and until the first half of the 20th century it was dominant in rural Syria (it is possible that it existed also in Aleppo in the early 20th century, if كامل الغزي is to be believed).

§ο. In some northern Egyptian dialects (in Cairo, Alexandria, and along the Damietta branch of the Nile) ق is pronounced ʔ, i.e. a hamza·t or glottal stop. In the rest of the Egyptian dialects ق is pronounced g. There is in northern Egypt a correlation between the pronunciation of ج and ق , the dialects which pronounce the first g pronounce the second ʔ, and the dialects which pronounce the first ǵ pronounce the second g.

The pronunciation ʔ for ق is found in pre-Hilālian dialects of the Maghreb. This pronunciation is also typical of the urban dialects of Greater Syria. According to Jean Cantineau (1938), this pronunciation existed in Aleppo, Latakia, Hama, Homs, Tripoli, Beirut, Damascus, Saida, Jerusalem, Hebron, etc., and it existed in rural dialects to the north of Damascus and in the northern half of Lebanon. However, most of the rural dialects around the mentioned cities had the pronunciation q (in rural central Palestine the fronted pronunciation k for ق existed, vide infra §II.21.A.f.ρ.). Nowadays the pronunciation ʔ for ق has expanded in rural areas, e.g. today nobody in Lebanon pronounces q for ق except the Druze. Also much of the countryside around Aleppo has shifted to this pronunciation (in modern western Syria the pronunciation q is commonly associated with the Nuṣayris, although a significant number of Sunni Muslims in western Syria still retain this pronunciation).

In Cairo and Palestine ق is pronounced as a plain glottal stop ʔ, but in the region of Aleppo it is pronounced with emphasis ʔ̣. Lebanon and Damascus have a middle pronunciation. In the region of Aleppo ʔ̣ < q may sometimes be dropped like an original hamza·t, e.g. the word baqara·tuṋ بَقَرَةٌ “cow” may be pronounced bāra·t in Al-Bȃb.

§π. The geographical distribution of the pronunciation ʔ for ق points to a Himyaritic/Yemeni origin, although this pronunciation seems to be attested once in a Ṣafaitic inscription, where the word qyð̣ “(he) spent the dry season” (perhaps for qȃyað̣a قَاْيَظَ ) was spelt ʔyṣ₂ (this also shows a confusion of ð̣ and ṣ₂; see M. C. Macdonald in The Cambridge Encyclopedia of the World’s Ancient Languages). As far as I know, the pronunciation ʔ for ق has not been recorded in Yemen, but the analogous pronunciations ʕ for ق and ʔ̣ for غ are known in southern Yemen (the first is said to exist in ʔAbyan, the second in Yȃfiʕ and Ḍȃliʕ, where Classical ɣanamuṋ غَنَمٌ “sheep” and maɣribuṋ مَغْرِبٌ “west” are pronounced ʔ̣aṇaṃ and ṃaʔ̣ṛeḅ; the underdots represent velarization; see Martine Vanhove, Yemen in Encyclopedia of Arabic Language and Linguistics). It would seem that all of these phonetic changes are based on two basic changes ɢ > ʔ and ʁ > ʕ. For example, the pronunciation ʔ for ق could have begun from *ɢ (voiced uvular plosive) and this would have been retracted to a glottal stop ʔ. An indirect evidence for this hypothesis is the presence in modern western Syrian dialects of words with ʁ/ɣ for etymological ق , e.g. *ɢidir > *ɢiḍir > ʁəḍer/ɣəḍer for qadira قَدِرَ “(he) be(came) able” (this word is found in rural dialects to the west of Aleppo; some people in Aleppo use it). The Aramaic change *ṣ₂ > ʕ could have happened in the following steps *ṣ₂ > (*ð̣ >) *ʁ > ʕ. In Old Aramaic the sound *ʁ (voiced uvular fricative) was written <q>, perhaps because it alternated with . This is supported by cuneiform writings in which Old Aramaic *ṣ₂ was written both <ḫ> (perhaps for ) and <q> (perhaps for ), e.g. Ra-ḫi-a-nu /*Raʁyān/ and Ra-qi-a-nu /*Raɢyān/ for *Raṣ₂yān “pleased.” The Aramaic development *ṣ₂ > (*ð̣ >) *ʁ would suggest that the Aramaic emphatics were uvularized (there are other indications for uvularization or pharyngealization in Aramaic, like the regular change *ḳt > ḳṭ, etc.).



Compare also the change in the pronunciation of Ancient Egyptian hieroglyph ꜣ , which is thought by some (e.g. Antonio Loprieno (1995), Ancient Egyptian: A linguistic introduction, p. 33) to have been pronounced ʀ (uvular trill) in the period of the Old Kingdom, but this was retracted to a glottal stop ʔ in the period of the Middle Kingdom.

Some Egyptologists (the followers of Otto Rössler or the “neuere Komparatistik” school) believe that ꜥ hieroglyph ꜥ (which sounded ʕ ) corresponds to Semitic d, ð and z (examples are Egyptian ꜥr.t /ʕrt/ “portal” vs. Semitic *dalt– “door,” and Egyptian ꜥffỉ /ʕffʔ/ “fly” vs Semitic ðbb “fly” (Akkadian zumbu, Arabic ðubȃbuṋ ذُبَاْبٌ )). If this correspondence is correct, it shows another instance of coronal consonants jumping back to the throat, although the details of this particular jump are mysterious.


§ρ. The pronunciation k for ق exists in rural central Palestine, in Suxna·t, in the dialect of the rural Baħȃrina·t الْبَحَاْرِنَةُ of eastern Arabia, and in some villages on the western and southern slopes of the Green Mountain of Oman. It should be noted that the dialects which have this pronunciation have also an unconditionally fronted pronunciation ᵗš for ك . This would suggest that the change *ḳ > k ق in those dialects motivated the change *k > ᵗš ك . By analogy, one may presume that the Bedouin change *ġ > g ق (where ġ means an emphatic voiced velar stop) was the motivation behind the widespread change *g > ᵈž ج .

§σ. The dialects which conditionally palatalize k ك sometimes also palatalize q ق . This is seen in rural gələt dialects of Mesopotamia, e.g. ᵈžarya·t for Classical qarya·tuṋ قَرْيَةٌ “village,” and θiᵈžīl for Classical θaqȋluṋ ثَقِيْلٌ “heavy.” A similar palatalization is seen in Kuwait, e.g. bāyi for Classical bȃqiṋ بَاْقٍ “remainder,” and riyy for Classical riyquṋ رِيْقٌ “saliva.”

►II.21.A.g. THE NASALS AND LIQUIDS m, n, l, r

§α. The sound m م is sometimes denasalized to b ب for no obvious reason, e.g. ħabilat حَبِلَتْ “she became pregnant” vs. ħamalat حَمَلَتْ “she bore;” and Bakka·tu بَكَّةُ a variant of Makka·tu مَكَّةُ “Mecca.” In modern Egyptian Arabic bitāʕ  بِتَاْعْ “thing” from mitȃʕuṋ مِتَاْعٌ .



The name of Aleppo was apparently written Ḫalam /Xalam/ in Ebla tablets (from the 3rd millennium BC), but in the Old Babylonian period (2nd millennium BC) it was Ḫalab /Xalab/. In the Neo-Assyrian period (1st millennium BC) the older name surfaced again as Ḫalman /Xalmān/. The current name Ħalabu حَلَبُ reflects the merger χ > ħ which was an areal phenomenon of the Greater Syrian region in the last millennium BC.


§β. In some dialects of Classical Arabic the sound n ن was apparently unstable in the syllable-terminal position. In Taǵwȋd (rules for perfect pronunciation of the Qurʔȃn) reciters are instructed to pronounce n ن as a nasal vowel whenever it is in syllable-terminal position and followed by a coronal obstruent or a velar/uvular stop. This rule is called الْإِِخْفَاْءُ “hiding (of syllable-terminal n).” This is probably Sibawayh’s “light nȗn” اَلْنُوْنُ اَلْخَفِيْفَةُ which he put on the top of his list of the good accessory letters (§II.20.A.b.). Perhaps this pronunciation of n explains such forms as muð مُذْ for munðu مُنْذُ “since” and lam yaku لَمْ يَكُ for lam yakun لَمْ يَكُنْ “[he] was not.” It seems that a syllable-terminal n was sometimes changed before coronals to ɲ and then to y. This produced ʔiysȃnuṋ إِيْسَاْنٌ for ʔinsȃnuṋ إِنْسَاْنٌ “human being,” ṣaydalȃniyyuṋ صَيْدَلْاَنِيٌّ for ṣandalȃniyyuṋ صَنْدَلَاْنِيٌّ “pharmacist” (from ṣandaluṋ صَنْدَلٌ “sandalwood”), and modern forms like mšān “for” < mišān < miyšān < minšān < min šaʔni مِنْ شَأْنِ . (See also Chaim Rabin p. 146 §qq. His explanations are probably incorrect.)

§γ. A useful thing to know about Arabic (and Afro-Asiatic) phonology is that the sounds n, l, r can alternate relatively easily.

In Aleppo (and other Syrian dialects) a pair of pants is called banṭarōn, from the European pantalon. The word bal “perhaps” (from Persian بل كه ) is pronounced barkī. The expression layta يَاْ لَيْتَ “I wish” is pronounced rēt. The word rāyeħ “going,” which is used to form the future tense, can be reduced in Aleppo to rāħ (e.g. rāħ yākol “[he] will eat”); this is pronounced in other Syrian regions l (e.g. laħ yākol “he will eat”).

The word yalʕanu يَلْعَنُ “(he) curses/will curse” is pronounced in some modern dialects yinʕan. The word malȋħuṋ مَلِيْحٌ “handsome” is pronounced in Aleppo (and other Syrian dialects) mnīħ. The word burtuqȃluṋ بُرْتُقَاْلٌ “oranges” (from Italian Portogallo) is pronounced in some Syrian dialects burdqān. The male personal name ᵈŽibrȋlu جِبْرِيْلُ “Gabriel” is pronounced in Sudan ᵈŽibrīn. The Persian word sangīn سنگين “adorned, rich” is pronounced in some dialects zangīl. The pronoun niħnā “we” is pronounced in some western Syrian dialects liħna or riħna; the pronunciation laħna is found in some dialects of ʔAbyan, and raħna is found in Yȃfiʕ. The 1st person plural verbal prefix n– is changed to l– in some dialects of ʔAbyan (e.g. lǝnṣa “(we) forget”) and in Γayl Ħabbȃn (e.g. lilʕab “(we) play”).

According to Ǵawȃd ʕAliyy (in المفصل في تاريخ العرب قبل الإسلام , vol. VIII, p. 574, quoting ابن الأثير ), the word la-ʕalla-ka لَعَلَّكَ “hopefully you” had the dialectal variants la-ʔanna-ka لَأَنَّكَ , la-ʕanna-ka لَعَنَّكَ , ra-ʕanna-ka رَعَنَّكَ , and ra-ɣanna-ka رَغَنَّكَ .


اترك تعليقًا

إملأ الحقول أدناه بالمعلومات المناسبة أو إضغط على إحدى الأيقونات لتسجيل الدخول:

شعار ووردبريس.كوم

أنت تعلق بإستخدام حساب تسجيل خروج   /  تغيير )

Facebook photo

أنت تعلق بإستخدام حساب Facebook. تسجيل خروج   /  تغيير )

Connecting to %s