►II.21.A.d. THE VELAR STOPS k, ǵ
§α. Arabic originally has two velar stops k ك and g ج . Sibawayh’s description of ك is of a voiceless velar stop k (§II.20.B.5.). However, his description of ج indicates a fronted sound (§II.20.B.6.), perhaps a voiced palatal plosive ɟ.
§β. In some modern dialects k ك is palatalized near front vowels to ᵗš = t͡ʃ or ᵗs = t͡s. The palatalization k > ᵗš is typical for eastern Arabia and the gələt dialects of Mesopotamia. The palatalization k > ᵗs is typical for Naǵd.
For example, the phrase la-ki لَكِ “for you(fem. sing.)” is pronounced li-ᵗš or li-ᵗs. The word bȃkiruṋ بَاْكِرٌ “early, morrow” is pronounced bȃᵗšir or bȃᵗsir. The word kalbuṋ كَلْبٌ “dog” is pronounced ᵗšalib or ᵗsalib. The word kam كَمْ “how many, how much” is pronounced ᵗšam or ᵗsam.
In most Yemeni dialects the change ki > š is found in the 2nd person feminine singular connected object pronoun –š “you(obj. fem. sing.).”
Sibawayh (in chapter 504) appears to have described these phenomena:
فَأَمَّا نَاسٌ كَثِيرٌ مِنْ تَمِيمَ وَنَاسٌ مِنْ أَسَدٍ فَإِنَّهُمْ يَجْعَلُونَ مَكَانَ الْكَافِ لِلْمُؤَنَّثِ الْشِّينَ، وَذَلِكَ أَنَّهُمْ أَرَادُوا الْبَيَانَ فِي الْوَقْفِ، لِأَنَّهَا سَاكِنَةٌ فِي الْوَقْفِ فَأَرَادُوا أَنْ يَفْصِلُوا بَيْنَ الْمُذَكَّرِ وَالْمُؤَنَّثِ، وَأَرَادُوا الْتَّحْقِيقَ وَالتَّوْكِيدَ فِي الْفَصْلِ لِأَنَّهُمْ إِذَا فَصَلُوا بَيْنَ الْمُذَكَّرِ وَالْمُؤَنَّثِ بِحَرْفٍ كَانَ أَقْوَى مِنْ أَنْ يَفْصِلُوا بِحَرَكَةٍ […] وَذَلِكَ قَوْلُكَ “إِنّشْ ذَاهِبَةٌ” وَ”مَالشْ ذَاهِبَةً”، تُرِيدُ “إِنَّكِ” وَ”مَالَكِ”. [كتاب سيبويه، باب الكاف التي هي علامة المضمر]
Sibawayh says that many of the Tamȋm (in eastern Arabia) and ʔAsad (in northeastern Naǵd) replaced the pronoun –ki “you(fem. sing.)” with šȋn at pause. Since Sibawayh’s šȋn was apparently a palatal fricative ç (vide §II.21.A.c.β.VI.), the change which he describes could have been –ki > –ç (if we take what he says literally). This must be an antecedent of the modern change k > ᵗš of eastern Arabia and Mesopotamia. The change k > ᵗs is perhaps a further development of that (i.e. k > ᵗš > ᵗs).
According to تاج العروس (under ديش ), “people who turned kȃf into šȋn” said ʔ·ad-dȋšu اَلْدِّيْشُ for ʔ·ad-dȋku اَلْدِّيْكُ “the rooster.”
الدِّيشُ بِالْكَسْرِ الدِّيكُ، لُغَةٌ فِيهِ عِنْدَ مَنْ يَقْلِبُ الكَافَ شِينًا، شَبَّه كافَه بكافِ المُؤَنَّثِ لِكَسْرَتِهَا. وأَنْشَدَ ثَعْلَبٌ “وإِنْ تَكَلَّمْتِ حَثَتْ في فِيشِ … حَتَّى تَنِقِّي كنَقِيقِ الدِّيشِ”
Sibawayh (ibid.) described other phenomena that are seemingly related to the palatalization mentioned above, but in fact they may be of totally different nature. According to him, some unspecified Arabs would attach a sȋn or šȋn to the pronoun –ki at pause “in order to make the terminal vowel –i apparent” (لِيُبَيِّنُوا كَسْرَةَ الْتَّأْنِيثِ).
وَاِعْلَمْ أَنَّ نَاسًا مِنْ الْعَرَبِ يُلْحِقُونَ الْكَافَ الْسِّينَ لِيُبَيِّنُوا كَسْرَةَ الْتَّأْنِيثِ، وَإِنَّمَا أَلْحَقُوا الْسِّينَ لِأَنَّهَا قَدْ تَكُونُ مِنْ حُرُوفِ الزِّيَادَةِ فِي اِسْتَفْعَلَ، وَذَلِكَ “أَعْطَيْتُكِسْ” وَ”أُكْرِمُكِسْ”، فَإِذَا وَصَلُوا لَمْ يَجِيئُوا بِهَا، لِأَنَّ الْكَسْرَةَ تَبِينُ. وَقَوْمٌ يُلْحِقُونَ الْشِّينَ لِيُبَيِّنُوا بِهَا الْكَسْرَةَ فِي الْوَقْفِ، كَمَا أَبْدَلُوهَا مَكَانَهَا لِلْبَيَانِ، وَذَلِكَ قَوْلُهُمْ “أَعْطَيْتُكِشْ” وَ”أُكْرِمُكِشْ”، فَإِذَا وَصَلُوا تَرَكُوهَا. [كتاب سيبويه، باب الكاف التي هي علامة المضمر]
It appears that what Sibawayh describes here is a phenomenon that belongs to §II.19.C., that is, he appears to say that terminal –ki was changed at pause to –kis or –kiš; both of these are perhaps altered from pausal *–kī. If this is what he says, then this change has nothing to do with palatalization but it is another variant of the arbitrary change –CVV > –CVC.
Other linguists dubbed the change –ki > –š (–ç) the “šanšana·t” شَنْشَنَةٌ , the change –ki > –kis the “kaskasa·t” كَسْكَسَةٌ , and –ki > –kiš the “kaškaša·t” كَشْكَشَةٌ . As for the location of the kaskasa·t and kaškaša·t, it seems that most of the linguists after Sibawayh located those phenomena in eastern Arabia (in the Rabȋʕa·tu رَبِيْعَةُ , which is a general designation for the eastern Arabian tribes).
A complete, unconditioned change k > ᵗš is found 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 (see also §II.21.A.f.ρ.).
§γ. The original sound of the voiced velar stop g ج is found today in the southwestern corner of Arabia (in Yemen from the Tihȃma·t coast up to ʔAbyan), in the southeastern corner of Arabia (in the mountainous northeast of Oman), and in northern Egypt (in Cairo, Alexandria, and along the Damietta branch of the Nile).
However, it seems that the pronunciation g for ج existed formerly in Iraq. The linguist ʔIbn Yaʕȋš اِبْنُ يَعِيْشَ , who lived in the 12-13th centuries AD, said that this pronunciation was common in Baghdad in his time (vide infra §II.21.A.a.δ.IV.). ʔ·Al-Maʕarriyy, who lived in northern Syria in the 10th-11th centuries AD, implied in his fictional رِسَالَةُ الْغُفْرَانِ that the pronunciation g for ج was typical of the ʕIbȃd of Ħȋra·tعِبَاْدُ الْحِيْرَةِ (the oldest inhabitants of Ħȋra·t). In his fictional story, which takes place in the afterlife, ʔ·Al-Maʕarriyy makes the 6th century AD ʕIbȃdi poet عَدِيٌّ بْنُ زَيْدٍ الْعِبَادِيُّ pronounce ج like ك (by which is probably meant the Persian g گ ).
فَيَقُولُ عَدِيٌّ بِعِبَادِيَّتِهِ “يَاْ مَكْبُوْر، لَقَدْ رُزِقْتَ مَاْ يَكِبُ أَنْ يَشْغَلَكَ عَنْ الْقَرِيْضِ، إِنَّمَا يَنْبَغِي أَنْ تَكُونَ كَمَا قِيْلَ لَكَ: كُلُوْا وَاِشْرَبُوْا هَنِيْئًا بِمَاْ كُنْتُمْ تَعْمَلُوْنَ”. قَوْلُهُ “يَاْ مَكْبُوْر” يُرِيدُ “يَاْ مَجْبُوْر”، فَجَعَلَ الْجِّيمَ كَافًا، وَهِيَ لُغَةٌ رَدِيئَةٌ يَسْتَعْمِلُهَا أَهْلُ الْيَمَنِ. وَجَاءَ فِي بَعْضِ الْأَحَادِيثِ أَنَّ الْحَارِثَ بْنَ هَانِىءٍ بْنَ أَبِي شَمِرِ بْنَ جَبَلَةَ الكِنْدِيِّ [تُوُفِّيَ فِي الْقَرْنِ الْمِيلَادِيِّ السَّابِعِ] اُسْتُلْحِمَ يَوْمَ سَابَاطَ [يَوْمٌ بِالْعِرَاقِ لَمَّا سَارْ سَعْدٌ مِنْ الْقَادِسِيَّةِ إِلَى الْمَدَائِنِ] فَنَادَى “يَا حُكْر يَا حُكْر” يُرِيدُ “يَا حُجْرَ بْنَ عَدِيٍّ الْأَدْبَرَ”. فَعَطَفَ عَلْيْهِ فَاِسْتَنْقَذَهُ. وَ”يَكِبُ” فِي مَعْنَى “يَجِبُ”. [المعري، رسالة الغفران]
And then ʕAadiyy says in his ʕIbȃdi [language], “yȃ makbȗru [yȃ magbȗru] يَاْ مَكْبُوْر , la-qad ruziqta mȃ yakibu [yagibu] يَكِبُ ʔan yašɣala-ka ʕan ʔ·al-qarȋḍi” (O magbȗru, you have been given things [in Paradise] which should keep you busy from poetry). You should do like you have been told [by God], “Eat and drink. Enjoy the reward for your [good] deeds.” By saying “yȃ makbȗru [yȃ magbȗru]” يَاْ مَكْبُوْر he wanted [to say] “yȃ maǵbȗru” يَاْ مَجْبُوْر . He made the ǵȋm a kȃf, which is a bad language used by the people of Yemen. It is told in some traditions that الْحَارِثُ بْنُ هَانِىءٍ بْنُ أَبِي شَمِرِ بْنُ جَبَلَةَ الكِنْدِيِّ [died in the 7th century AD] was surrounded by enemies during the Battle of Sȃbȃṭ [during the Islamic conquest of Iraq], and he cried “O Ħukr, O Ħukr [O Ħugr, O Ħugr]” يَا حُكْر يَا حُكْر , by which he wanted [to say] “O Ħuǵru ʔ·ibnu ʕAadiyyiṋ ʔ·al-ʔAdbaru” يَا حُجْرَ بْنَ عَدِيٍّ الْأَدْبَرَ . He [Ħuǵr] felt sympathy for him and salvaged him. As for yakibu [yagibu] يَكِبُ , it means yaǵibu يَجِبُ [“it is necessary”].
Arabic ج rendered Middle Iranian g in early Classical Arabic words, e.g. fȃlȗðaǵuṋ فَاْلُوْذَجٌ “sweetmeat made of flour, water, and honey” from Middle Persian pȃlȗdag, firǵȃruṋ فِرْجَاْرٌ “pair of compasses” from Middle Persian pargȃr, ṣawlaǵȃnuṋ صَوْلَجَانٌ “scepter” from Middle Iranian ᵗšawgȃn–, ṣanǵuṋ صَنْجٌ “harp” from Middle Persian ᵗšang, and šaṭranǵuṋ شَطْرَنْجٌ “chess” from Middle Persian ᵗšatrang.
§δ. One of the bad supplementary letters mentioned by Sibawayh (§II.20.A.c.) is the “ǵȋm which is like kȃf” اَلْجِيْمُ اَلَّتِيْ كَاَلْكَاْفِ . This probably refers to the pronunciation g for ج . The “kȃf which is between ǵȋm and kȃf” اَلْكَاْفُ اَلَّتِيْ بَيْنَ اَلْجِيْمِ وَاَلْكَاْفِ is likely a voiced allophone of k ك , i.e. it is the same sound g. This would explain why Sibawayh appears to have counted both of these two supplementary letters as one. This is the opinion of ʔIbn Yaʕȋš (or perhaps ʔIbn Durayd?)
فَأَمَّا الْكَافُ الَّتِي بَيْنَ الْجِيْمِ وَالْكَافِ، فَقَالَ اِبْنُ دُرَيْد: هِيَ لُغَةٌ فِي الْيَمَنِ، يَقُولُونَ فِي “جَمَلٍ” “كَمَلٌ”، وَفِي “رَجُلٍ” “رَكُلٌ”. وَهِيَ فِي عَوَامِّ أَهْلِ بَغْدَادَ فَاشِيَةٌ شَبِيهَةٌ بِالْلُّثْغَةِ. وَالْجِيمُ الَّتِي كَالْكَافِ كَذَلِكَ، وَهُمَا جَمِيعًا شَيْءٌ وَاحِدٌ، إِلَّا أَنَّ أَصْلَ إِحْدَاهُمَا الْجِيمُ، وَأَصْلَ الْأُخْرَى الْكَافُ، ثُمَّ يَقْلِبُونَهُمَا إِلَى هَذَا الْحَرْفِ الَّذِي بَيْنَهُمَا. [شرح المفصل لابن يعيش. فصل حروف العربية]
As for the kȃf which is between ǵȋm and kȃf, ʔIbn Durayd said, it is a language in Yemen, they say kamaluṋ [gamaluṋ] كَمَلٌ for ǵamaluṋ جَمَلٌ [“camel”], rakuluṋ [raguluṋ] رَكُلٌ for raǵuluṋ رَجُلٌ [“man”]. It is widespread among the common people of Baghdad, like a lisp. The ǵȋm which is like kȃf is the same. They are both the same thing, but one of them originates from ǵȋm and the other from kȃf, by turning both of them into this letter which is in between.
By saying that the pronunciation g for ج “is like a lisp” شَبِيْهَةٌ بِاَلْلُّثْغَةِ he perhaps meant to say that the people in Baghdad could not but speak in this incorrect manner.
A voiced allophone of k ك is found in a few words in Aleppo Arabic, e.g. ragad for rakaḍa رَكَضَ “(he) ran.” (The final *ṣ₂ > *ð̣ > *ð > d in the Aleppine word may be an isolated development, or it may be a remnant from a complete merger ṣ₂ > d that possibly existed in some early Middle Arabic dialects, see Lipiński ibid. p. 121.)
§ε. Some modern scholars have suggested that the sound of the correct ǵȋm described by Sibawayh is gʸ. This suggestion is based primarily on theoretical grounds, since it can explain why ج has been palatalized in most of the dialects. However, it does not seem very likely that gʸ was the sound of ǵȋm meant by Sibawayh and other linguists, because gʸ is too close to g, which was considered a bad sound by Sibawayh and others.
Sibawayh characterized ǵȋm literally as a voiced palatal plosive, and it could have been really a voiced palatal plosive ɟ. However, as far as I know there is no direct evidence for this. In the 8th century AD bilingual Psalm fragment from Syria (§II.21.A.c.β.VI.) Arabic ǵȋm was transcribed with Γ, which in the Greek of the 8th century AD had the sounds ɣ and ʝ, and several centuries earlier it had the sound g.
The old linguists talked about a dialectal change of terminal –iy(y) to –iǵ(ǵ), which they called ʕaǵʕaǵa·tuṋ عَجْعَجَةٌ . From the examples they gave for this phenomenon, it would seem that it was a pausal phenomenon that belongs to §II.19.C.
يَاْ رَبِّ إِنْ كُنْتَ قَبِلْتَ حَجَّتِجْ (حَجَّتِيْ)
فَلَاْ يَزَاْلُ شَاْحِجٌ يَأْتِيْكَ بِجْ (بِيْ)
أَقْمَرُ نَهَّاْزٌ يُنَزِّيْ وَفْرَتِجْ (وَفْرَتِيْ)
*********
خَاْلِيْ عُوَيْفٌ وَأَبُوْ عَلِجّ (عَلِيٍّ)
اَلْمُطْعِمَاْنِ اَلْلَّحْمَ بِاَلْعَشِجّ (بِاَلْعَشِيِّ)
وَبِاَلْغَدَاةِ كِسَرَ اَلْبَرْنِجّ (البَرْنِيِّ)
يَقْلعُ بِاَلْوَدِّ وَبِاَلْصِّيْصِجّ (اَلْصِّيصِيِّ)
*********
وَأَمَّا نَاسٌ مِنْ بَنِي سَعْدٍ فَإِنَّهُمْ يُبْدِلُونَ الْجِيمَ مَكَانَ الْيَاءِ فِيْ الْوَقْفِ، لِأَنَّهَا خَفِيَّةٌ فَأَبْدَلُوا مِنْ مَوْضِعِهَا أَبْيَنَ الْحُرُوفِ، وَذَلِكَ قَوْلُهُمْ ”هَذَا تَمِيمِجّ” يُرِيدُونَ ”تَمِيمِيٌّ”، و “هَذَا عَلِجَ” يُرِيدُونَ ”عَلِيٌّ”. وَسَمِعْتُ بَعْضَهُمْ يَقُولُ “عَرَبَانِجّ” يُرِيدُ “عَرَبَانِيٌّ”. وَحَدَّثَنِي مَنْ سَمِعَهُمْ يَقُولُونَ “خَالِي عُوَيْفٌ وَأَبُو عَلِجّ… الْمُطْعِمَانِ الشَّحْمَ بِالْعَشِجّ… وَبِالْْغَدَاةِ فِلَقَ الْبَرْنِجّ” يُرِيدُ “بِالْعَشِيِّ” و”الْبَرْنِيِّ”، فَزَعَمَ أَنَّهُمْ أَنْشَدُوهُ هَكَذَا. [كتاب سيبويه، باب الحرف الذي تبدل مكانه في الوقف حرفا أبين منه]
This phenomenon would be strange if we though of ǵȋm as gʸ or ᵈž, but if we accept that ǵȋm was pronounced ɟ (which is what Sibawayh literally says), it will turn out to be trivial. It is not any harder to imagine than the other pausal phenomena described in §II.19.C. (it is basically a change Cȋ > Ciɟ).
Jean Cantineau (in Etudes sur quelques parlers de nomades arabes d’Orient (1936)) mentioned the existence of the pronunciation ɟ for ج in northern Arabian Bedouin dialects of the Ruwala, ʕAneze, and Šammar. In those dialects the pronunciation ɟ alternated freely with g.
§ζ. Most of the modern dialects have palatalized pronunciations for ج in all environments. The most common pronunciation is ᵈž = d͡ʒ (a voiced post-alveolar affricate). The is realized in southern Syria and Lebanon as a weak affricate, while in Mesopotamia (and most other dialects) it is a strong affricate. In Bedouinic Šammar dialects in the Arabian Peninsula, Jordan, and Sinai it is realized dʸ.
The Arabic name of Algeria ʔ·Al-Ǵazȃʔiru الْجَزَاْئِرُ “the Islands” is pronounced locally Dzȃyir, from older *ᵈŽzȃyir جْزَاْيِرْ* with partial assimilation of the first consonant to the second one. A similar development produced the Algerian word bezzȃf “a lot, much” from bi-ʔ·al-ᵈžizȃfi بِاَلْجِزَاْفِ .
N.B. A common myth found in many modern books is that ج is pronounced ž = ʒ in Syria or Lebanon. This is not a usual pronunciation neither in Syria nor in Lebanon.
§η. An unaffricated pronunciation ž for ج is said to exist today in Samȃwa·t in southern Iraq, while in Baṣra·t and Kuwait ج is pronounced as a pure palatal y, e.g. in Baṣra·t yimal for ǵamaluṋ جَمَلٌ “camel,” and ʕayȗz for ʕaǵȗzuṋ عَجُوْزٌ “old [woman];” and in Kuwait yȃr for ǵȃruṋ جَاْرٌ “neighbor,” diyȃya·t for daǵȃǵa·tuṋ دَجَاْجَةٌ “a chicken,” and ya for ǵȃʔa جَاْءَ “(he) came.”
The pronunciation y for ج exists also in Ħaḍramawt and in the Bedouin dialects of western and southeastern Oman.
Strangely, the pronunciation ž for ج exists in Morocco, apparently by a local development g > ᵈž > ž, since there are allophones [g] and [d] that appear in words containing s or z, e.g. in ɡləs for ǵalasa جَلَسَ “(he) sat,” ɡəzzȃr for ǵazzȃruṋ جَزَّاْرٌ “butcher,” and ydȗz for yaǵȗzu يَجُوْزُ “(he) goes past/will go past.”
§θ. One of the bad supplementary Arabic letters mentioned by Sibawayh is the “ǵȋm which is like šȋn” اَلْجِيْمُ اَلَّتِيْ كَاَلْشِّيْنِ (vide §II.20.A.c.). This is perhaps an allophone of ǵȋm which Sibawayh described in chapter 569 ↗. He gave for it the written example ʔašdaru أَشْدَرُ for ʔaǵdaru أَجْدَرُ “[who] deserves more.” This example shows that the allophone was not voiceless, since it appeared before the voiced d. Thus, what is meant is probably a fricativized allophone of the plosive ɟ. ʔIbn Yaʕȋš اِبْنُ يَعِيْشَ (died in 1245 AD) said that the “ǵȋm which is like šȋn” appeared mostly before the dentals d and t. Based on this we may presume that the meant sound was a fronted ž = ʒ. This would explain why Sibawayh counted this “bad letter” separately from the good “šȋn which is like ǵȋm” اَلْشِّيْنُ اَلَّتِيْ كَاَلْجِيْمِ (§II.21.A.c.β.VIII.), which perhaps was a mere voiced allophone of šȋn, i.e. ʝ.
وَأَمَّا الْجِيمُ الَّتِي كَالْشِّينِ، فَهِيَ تَكْثُرُ فِي الْجِيمِ الْسَّاكِنَةِ إِذَا كَانَ بَعْدَهَا دَالٌ أَوْ تَاءٌ، نَحْوَ قَوْلِهِمْ فِي “اِجْتَمَعُوا” وَ”الْأَجْدَرُ” “اِشْتَمَعُوا” وَ”الْأَشْدَرُ”، فَتُقَرَّبُ الْجِيمُ مِنْ الْشِّينِ، لِأَنَّهُمَا مِنْ مَخْرَجٍ وَاحِدٍ، إِلَّا أَنَّ الْشِّينَ أَبْيَنُ وَأَفْشَى. فَإِنْ قِيلَ: فَمَا الْفَرْقُ بَيْنَ الْشِّينِ الَّتِي كَالْجِيمِ حَتَّى جُعِلْتَ فِي الْحُرُوفِ الْمُسْتَحْسَنَةِ، وَبَيْنَ الْجِيمِ الَّتِي كَالْشِّينِ حَتَّى جُعِلْتَ فِي الْحُرُوفِ الْمُسْتَهْجَنَةِ؟ قِيلَ: إِنَّ الْأَوَّلَ كُرِهَ فِيهِ الْجَمْعُ بَيْنَ الْشِّينِ وَالْدَّالِ لِمَا بَيْنَهُمَا مِنْ الْتَّبَايُنِ الَّذِي ذَكَرْنَاهُ؛ وَأَمَّا إِذَا كَانَتْ الْجِيمُ مُقَدَّمَةَ كَـ”الْأَجْدَرِ” وَ”اِجْتَمَعُوا” فَلَيْسَ بَيْنَ الْجِيمِ وَالْدَّالِ مِنْ الْتَّنَافِي وَالْتَّبَاعُدِ مَا بَيْنَ الْشِّينِ وَالْدَّالِ؛ فَلِذَلِكَ حَسُنَ الْأَوَّلُ وَضَعُفَ الْثَّانِي. [شرح المفصل لابن يعيش. فصل حروف العربية]
One of the examples given by ʔIbn Yaʕȋš for the “ǵȋm which is like šȋn” is ʔ·ištamaʕȗ اِشْتَمَعُوْا for ʔ·iǵtamaʕȗ اِجْتَمَعُوْا “(they) met, gathered.” This example is different from Sibawayh’s examples given in chapter 569 ↗, in which the tȃʔ is assimilated to the preceding ǵȋm, e.g. ʔ·iǵdamaʕȗ اِجْدَمَعُوْا for ʔ·iǵtamaʕȗ اِجْتَمَعُوْا “(they) met, gathered.” However, this difference probably does not mean that ʔIbn Yaʕȋš is mistaken.
§ι. ʔIbn ʔ·al-Ǵazariyy اِبْنُ اَلْجَزَرِيِّ (died in 1429 AD) described a sound change similar to that described by ʔIbn Yaʕȋš and (very vaguely) by Sibawayh. According to him, some people would pronounce ǵȋm incorrectly (يَغْلَطُونَ فِيهَا) when it is followed by zȃy or sȋn; they would give it hamsuṋ هَمْسٌ (i.e. the quality of mahmȗs §II.20.C.) and raxȃwa·tuṋ رَخَاْوَةٌ (i.e. the quality of raxw §II.20.D.). However, in some of the examples which he mentioned there is a dental plosive after the ǵȋm.
وَإِذَا سَكَنَتْ الْجِيمُ، سَوَاءً كَانَ سُكُونُهَا لَازِمًا أَمْ عَارِضًا، فَإِنْ كَانَ لَازِمًا وَجَبَ الْتَّحَفُّظُ مَنْ أَنْ تَجْعَلَ شِينًا، لِأَنَّهُمَا مِنْ مَخْرَجٍ وَاحِدٍ، فَإِنَّ قَوْمًا يَغْلَطُونَ فِيهَا، لَا سِيَّمَا إِذَا أَتَى بَعْدَهَا زَايٌ أَوْ سِينٌ، فَيُحْدِثُونَ هَمْسًا وَرَخَاوَةً وَيُدْغِمُونَهَا فِي الْزَّايِ وَالْسِّينِ وَيُذْهِبُونَ لَفْظَهَا، وَذَلِكَ نَحْوَ قَوْلِهِ “اِجْتَمَعُوا” وَ”الْنَّجْدَيْنِ” وَ”اِجْتَنبُوا” وَ”خرجت” وَ”وَجْهَكَ” وَ”تَجْزِي” وَ”تُجْزَوْنَ” وَ”رجزًا” وَنَحْوَ ذَلِكَ، فَلَا بُدَّ أَنْ يُنْطَقَ بِجَهْرِهَا وَشِدَّتِهَا وَقَلْقَلَتِهَا. وَإِذَا كَانَ سُكُونُهَا عَارِضًا فَلَا بُدَّ مِنْ إِظْهَارِ جَهْرِهَا وَشِدَّتِهَا وَقَلْقَلَتِهَا، وَإِلَّا ضَعُفَتْ وَاِنْمَزَجَتْ بِالْشِّينِ، وَذَلِكَ نَحْوَ قَوْلِهِ “أُجَاجٌ” وَ”فخراج” وَنَحْوَ ذَلِكَ فِي الوَقْفِ. [ابن الجزري، التمهيد في علم التجويد]
What ʔIbn ʔ·al-ᵈŽazariyy says is that ǵȋm incorrectly becomes mahmȗs and raxw (fricative) before coronal consonants, whereas the correct pronunciation should be maǵhȗr, šadȋd (plosive), and with qalqala·tuṋ قَلْقَلَةٌ , which means the insertion of a schwa mobile after the ǵȋm. Despite using the word mahmȗs, the incorrect pronunciation which he means is likely ž = ʒ. He might have used the word mahmȗs because a fricative is naturally accompanied by more breath than a plosive (especially in Arabic). It is also possible that ž was devoiced before voiceless coronals but he generalized this. Since ʔIbn ʔ·al-ᵈŽazariyy lived in the 14th-15th centuries, the correct pronunciation of ǵȋm which he had in mind could have been ᵈž rather than Sibawayh’s ɟ.
§κ. The pronunciation ᵈž/ž for ج may be ancient. The ANA inscriptions from Ħismȃ (the sand-desert of southern Jordan and northwest Saudi Arabia) used the character θ 
to write etymological g. This perhaps indicates the pronunciation ᵈž or ž. On the other hand, Ṣafaitic g was used to write Greek γ and Latin g, e.g. Ṣafaitic Grgs₁ for Γεώργιος (Geṓrgios) and Ṣafaitic Grmnqṣ for Germanicus. This spelling perhaps shows that Ṣafaitic g was still hard.
The pronunciation ž for ج , despite being relentlessly denounced by the Muslim linguists and Qurʔȃn reciters, must have been common in Naǵd at some point, because this would explain why the pronunciation ž/y for ج is found today in southern Iraq, Kuwait, Ħaḍramawt, western and southeastern Oman, ᵈŽawf, and among certain north Arabian tribes (notably the Sardiyye and the Sirħȃn). The former existence of the pronunciation ž for ج in Cairo and Damascus may be inferred from such words as wišš (in Cairo) and wəšš (in Damascus) for Classical waǵhuṋ وَجْهٌ “face.” The word wišš/wəšš seems to reflect an earlier *wižh. The corresponding word in Aleppo and Lebanon is wəǵǵ. In the dialect of Damascus there is also tzawwaz for Classical tazawwaǵa تَزَوَّجَ “(he) married.” This seems to reflect an earlier *tzawwaž. Also in Damascus there is saᵈžara·t for Classical šaǵara·tuṋ شَجَرَةٌ “tree,” which is apparently the result of dissimilation *šažara·t > *sažara·t > saᵈžara·t. The former existence of words with ž for ج in Damascus and Cairo is perhaps due to Bedouinic influence from the Syrian Desert.
§λ. Benjamin Hary (and before him Haim Blanc) suggested that the pronunciation of ج in Cairo changed from g to ǵ in the 6th-7th centuries, and it changed back to g in 19th-20th centuries. This theory is based on Middle Arabic texts written in Hebrew characters from the Cairo Geniza, in which the character גֹ or גִ (gîmēl with a supralinear or sublinear dot) was used to indicate a front pronunciation of Arabic ج . However, as Liesbeth Zack notes:
[… C]aution is required when using the language of these Judeo-Arabic documents as proof of the pronunciation of Egyptian Arabic in general. First of all, it is not at all certain if the reported speech really reflects the dialect of Cairo. Secondly, it is not definitely known whether the Jews of Cairo spoke the same dialect as the Muslims at that time, or if there were any differences. What is, however, known is that the Jewish dialect in Cairo in the twentieth century differed in several ways from the dialect spoken by the muslims.
Liesbeth Zack (2009), Egyptian Arabic in the seventeenth century : a study and edition of Yȗsuf al-Maġribȋ’s Daf’ al-iṣr ‘an kalȃm ahl Miṣr, p. 85
Blanc’s theory was further supported by certain quotations and interpretations from the early 17th century book دَفْعُ الْإِصْرِ عَنْ كَلَامِ أْهْلِ مِصْرَ by يُوسُف الْمَغْرِبِي (died in 1611 AD). The author of this book said that [some of] the country people in Egypt (يُسْمَعُ من أهل الريف) pronounced the word raǵuluṋ رَجُلٌ “man” as raǵl رَجْل , with “a vowel a after the rȃʔ and no vowel after the non-Arabic ǵȋm” (بفتح الرا وسكون الجيم الغير العربية). His explanation of the “non-Arabic ǵȋm” shows that it was none but g.
In another passage the author says that he heard a man who was watering camels (a Bedouin?) say چوچو “come to drink.” He explains چ as the sound of “the non-Arabic ǵȋm which is close to šȋn” (الجيم الغير العربية القريبة للشين), by which he obviously means ž.
Since this author was from Cairo, what he said was taken as evidence that the pronunciations g and ž for ج were unusual in Cairo in the early 17th century AD.
However, this author never said that the pronunciation g for ج did not exist in Cairo. He only called this pronunciation “non-Arabic,” by which he doubtless meant non-standard Arabic, because the word “Arabic” for the older authors meant only the standard language which was defined by the early grammarians of Iraq. As Liesbeth Zack notes (ibid. p. 86), the author may have described the pronunciation of ǵȋm in this particular instance because in Egypt the rural dialects usually have the pronunciation ᵈž for ج .
One should note that this author (like other old authors) draws a line between the speech of the common people اَلْعَوَاْمُّ and the speech of the special people اَلْخَوَاْصُّ . This line must have been mostly an imaginary line, given the author’s claim (on the side of fol. 11a) that the people who say t for θ ث are the common people (المُحَدِّثِين بالتاء المثناة فوق أعني بهم العوام).
Since the author himself confused t ت for θ ث and d د for ð ذ many times in his book (see Liesbeth Zack ibid. p. 89), it is obvious that his own speech was not different from the speech of the common people. What he meant by the speech of the special people حديث الخواصّ is the literary standard language, i.e. an imaginary language that nobody really spoke. This author (like other old authors) renounced his own speech and considered himself a speaker of Classical Arabic.
A somewhat similar psychological stance led one author who died in 1933 to claim something that was almost completely detached from reality. This was كامل الغَزِّي who wrote a book named نهر الذهب في تاريخ حلب . In page 236 he talks about the dialect of Aleppo (لهجة أهل حلب في التكلم):
الحلبي يلفظ الحرف من مخرجه الحقيقي فيلفظ الجيم جيما لا كيما عبرانية ولا شينا ولا زايا، والشين شينا لا سينا، والقاف قافا لا كافا ولا همزة وهكذا بقية الحروف. غير أن النصارى واليهود يخرجون القاف همزة مرققة في الغالب فيقولون في قارىء مثالا آري. كما أن بعض النصارى ربما أخرجوا التاء طاء فقالوا في ترى مثلا طرى. أو أخرجوا السين صادا فقالوا في ساعة مثلا صاعة. لكن هذا قليل. وقد يخرج اليهود الضاد والطاء بين التاء والدال فقالوا في مثل فضله وأعطني فدله وأعتني. ويوجد بعض من المسلمين الذين يعاملون عرب البادية من يخرج القاف كافا مفخمة فيقولون في قال مثلا كال. والإعراب في كلامنا لا وجود له البتة. وتوجد فيه الإمالة بكثرة جائزة وممتنعة كقولهم سريج لحيف قيعد نييم في سراج لحاف قاعد نائم. […] ولا وجود للذال في كلامهم بل يقلبونها دالا أو يلفظون بها زايا، وتاء التأنيث في الأسماء يقلبونها ياء كقولهم فاطمي عاقلي أي فاطمة عاقلة، والثاء تاء كقولهم تيني تيلت أي ثاني ثالث، وربما لفظوها سينا كقولهم في ثم سم، ويخرجون الظاء زايا مفخمة كقولهم في ظاهر زاهر أو يقلبونها ضادا كقولهم أَدَّن الضهر في أَذَّن الظهر. وكثير من المتظرفين الذين يعانون اللغة التركية من يلفظ الضاد زايا مفخمة فيقول في مريض مثلا مريز. كما أن بعض المتفرنجين يلفظون القاف كافا فيقول في قرش مثلا كرش حتى إنهم كثيرا ما يخلطون في كلامهم العربي بعض كلمات افرنجية لا عجزا منهم عن أن يأتوا بنظائرها من اللغة العربية بل ليُفهموا أن معرفتهم باللغة العربية قليلة كأنهم يفتخرون بذلك. وبالجملة فإن لكل محلة عندنا لهجة مخصوصة بأهلها غلب عليهم ذلك باعتبار خلطائهم، فسكان محلة باب النيرب يغلب على لهجتهم ألفاظ أهل الصوف والوبر، وسكان محلة الكلاسة يغلب عليهم ألفاظ أهل الحرث، وسكان محلة الفرافرة يوجد في كلامهم كثير من الألفاظ التركية والفارسية إذ كان أكثر المستخدمين في أيام الحكومة العثمانية منهم، والنصارى واليهود يغلب عليهم في تراكيب كلامهم أساليب اللغات الافرنجية لكثرة معاناتهم إياها، وهكذا بقية المحلات، كُلٌّ يُضَارِعُ خَلِيطَهُ.
The Aleppine pronounces the letters at their true points of articulation. He pronounces the ǵȋm as a [true] ǵȋm, not as an Hebraic kȋm [gȋm], nor as a šȋn or zȃy, and [he pronounces] the šȋn as a [true] šȋn, not as a sȋn, and [he pronounces] the qȃf as a [true] qȃf, not as kȃf nor as a hamza·t, and so for the rest of the letters. However, the Nazarenes [Christians] and the Jews usually make the qȃf a soft hamza·t, for example, they say ʔȃrȋ آري for qȃriʔuṋ قارئ [“reading, reader”]. Some of the Nazarenes may also make the tȃʔ a ṭȃʔ, for example, they would say ṭara·y طرى for tara·y ترى [“you see”]. They may also make the sȋn a ṣȃd, for example, they say ṣȃʕa·t صاعة for sȃʕa·tuṋ ساعة [“hour”]. However, this is not common. The Jews may make the ḍȃd and ṭȃʔ between the tȃʔ and dȃl, they would say fadl-o [?] فدله and ʔaʕti-niy أعتني for faḍlu-hu [?] فضله [“his favor”?] and ʔaʕṭi-niy أعطني [“give me”]. Some of the Muslims who deal with the Arabs of the desert make the qȃf an amplified kȃf كاف مفخمة . For example, they say kȃl [ġȃl] كال for qȃla قال [“(he) said” …]
Although the author called this section “the dialect of the people of Aleppo,” the first part of the section is not very much a description of Aleppine phonology in comparison to Standard Arabic and other dialects, rather it seems that this part was meant to show that the Muslims of Aleppo spoke better Arabic than the Christians and Jews. The author begins by claiming that the [Muslim] people of Aleppo pronounced every letter correctly, and then he points out some deviations in the speech of the Christians and Jews. He claims that the Muslims pronounced ق in its correct sound q, while the Christians and Jews pronounced it as a hamza·t.
The famous Jean Cantineau published a description of Syrian dialects in the 1930’s, i.e. shortly after this book was published. In Cantineau’s description it is stated that ق was pronounced ʔ in Aleppo (just like in all the other urban dialects of Syria). As far as I know, Cantineau did not mention the existence of the pronunciation q for ق in Aleppo, although in his time this pronunciation was still dominant in the countryside around the city. Nowadays this pronunciation has largely disappeared from the western (and northern?) countryside and has been replaced by the urban pronunciation ʔ̣. I lived in Aleppo for several years and I never heard a native person pronounce the sound q in any native word (no matter how old he or she was). There are no traces of such a pronunciation in any layer of the dialect.
It is probable that the statement about the pronunciation of ق in Aleppo by كامل الغَزِّي was incorrect. His illusional statement could have had a subconscious religious motive behind it (given that he was a religious sheikh). I once listened to a voice-recording of an old Christian woman from Aleppo (it was published somewhere on the internet); her speech in the recording was rather unusual for a person from Aleppo, because she pronounced the etymological sounds aw and ay without the usual contractions to ō and ē. The person who was interviewing her asked her why she spoke like that. She answered “this is how the Christians speak.” The interviewer told her that he had never heard any Christian in Aleppo speak like her. She answered “they imitate the Muslims.” In my opinion, it is likely that كامل الغَزِّي had the same psychological stance. The pronunciation q for ق was probably his own personal pronunciation, but by a certain psychological process he became convinced that this was the Muslim pronunciation, whereas the pronunciation ʔ̣ was a Christian and Jewish pronunciation.
الراصد العربي 