Reconciling Contradictory Evidence on Quranic Arabic

 

I have written several English articles on wikiislam.net detailing the evidence for the Old Hijazi reconstruction of the original language of the Quran. I have also contributed to the strongest evidence for this reconstruction by discovering hundreds of additional internal rhymes that only appear if the Quran is rid of final short vowels and Tanwin.
The strength of the internal rhyme evidence makes it unquestionable that the Quran originally lacked full Iʕrāb. But there are still some issues that the Jallad-Putten model has no answer for.
The Jallad-Putten model can be summarized into the following points:
1- The Quran was composed in the vernacular of the Hijaz. This language, called Old Hijazi, lacked a hallmark of Classical Arabic and the reading traditions which is final short vowels and Tanwin (Iʕrāb).
2- The Uthmānic standard Quran was a phonetic text written in Old Hijazi. The mismatch between the standard text and the pronunciation is the result of the later introduction of Iʕrāb into the Quran.
3- The people of the Hijaz didn’t compose poetry in the classical form of QaṢīdah. But sometime during the Umayyad rule, the QaṢīdah register became the high register while Old Hijazi become only a performative register. Muslims then applied this new high register on the Quran by introducing Iʕrāb.

The second and third points can be challenged. To resolve these challenges, I propose that pre-Islamic Hijazis had two registers: A poetry register that had full Iʕrāb and a spoken vernacular that lacked Iʕrāb. The Quran was composed in the vernacular register. Sometime before the Uthmānic recension, Iʕrāb started to get adopted in melodic recitation. But in normal speech such as when quoting Quranic verses during a sermon, Iʕrāb wasn’t used. By the time the Uthmanic recension was compiled, reciting the Quran with Iʕrāb had already become standard practice, which explains why the oral ancestor of the canonical readings had Iʕrāb. But since that Iʕrāb wasn’t applied to the Quran in normal speech, Iʕrāb was dropped when the Quran was dictated to the Uthmanic scribes. This explains why the written text preserves clues, such as the treatment of the feminine ‘t’, pointing to an original Iʕrāb-less pronunciation. Iʕrāb was initially adopted in melodic recitation because it makes the text more melodic and easier to sing. Over time, however, applying Iʕrāb to the Quran became ideological which contributed to the complete disappearance of the original Iʕrābless pronunciation.


1- Iʕrāb was applied earlier.
The Jallad-Putten model suggests that Iʕrāb was applied during the Umayyad rule. But evidence shows that Iʕrāb was already applied at the time of the Uthmānic recension.

A- A unique reading proves Iʕrāb in the common ancestor of the canonical readings
Verse 53:50 ends with these two words:
﴿عَادًا ٱلۡأُولَى﴾
ʕādani l-ʾūlā
The four canonical readers Nāfiʕ, Abū jaʕfar, Abū ʕamr and yaʕqūb read ʕādani l-ʾūlā as ʕādall-ūlā. This form cannot arise if Tanwin wasn’t added to the word ʕād. According to the linguistic features of readings of previously mentioned four readers, the two words shouldn’t be pronounced as ʕādall-ūlā. But the four of them pronounced it as such. Plus, the 4 readers come from two different regions: Two from Median and two from Baṣrah. This means that the form ʕādall-ūlā was in the shared reading from which the canonical readings evolved. The fact that this unique form also results from the dropping of the Hamzah of “l-ʾūlā” solidifies that this form was the original reading since that the dialect of the people of Hijaz lacked Hamzah. All of this proves that Tanwin was part of the shared reading.

(The following part is a bit technical. A layman reader can skip it.)

The original form of the two words is: ʕād al-ʾūlā
The word ʕād is in the accusative case so it took the final short vowel ‘a’: ʕāda.
It’s also indefinite so it took a final ‘n’ (Tanwin) : ʕādan.
In Classical Arabic, the ‘a’ in the ‘al’ article is pronounced only when it comes at the beginning of a sentence. When the ‘al’ article comes in the middle of a sentence then the ‘a’ isn’t pronounced, leaving the ‘al’ article as a mere ‘l’. This yields the form:
ʕādan l-ʾūlā (instead of ʕādan al-ʾūlā)
Notice how the ‘n’ of ʕādan is followed by an L then a glottal stop. So we have here a cluster of three consonants: ‘n’, ‘l’ and the glottal stop. Classical Arabic rules dictate that a short vowel must be added to break up such clusters. In such cases, an ‘i' short vowel gets introduced to break up the cluster, which yields:
ʕādani l-ʾūlā (notice the added ‘i' at the end of ʕādan)
That’s how we went from the original form of ʕād al-ʾūlā to the Classical Arabic pronunciation of ʕādani l-ʾūlā.
In the dialect of Quraysh or the Hijaz region, the glottal stop of l-ʾūlā gets dropped: l-ūlā.
In this case, there’s no longer a cluster of consonants, and thus there’s no need to attach an ‘i' to the Tanwin to break up the cluster, so we get:

ʕādan l-ūlā.

Since that the n now is followed by an L, the two consonants can now merge into a stressed L:

ʕādall-ūlā

The form ʕādall-ūlā cannot arise if Tanwin wasn’t added to the word ʕād. According to the Jallad-Putten Model, the original pronunciation of the two words should be ʕādā lūlā.

B- Haytham Sidqi’s paper 
https://www.academia.edu/111507928/Consonantal_Dotting_and_the_Oral_Quran
This paper has proven that the canonical reading traditions share “an inherited oral tradition that likely dates to the time of the Uthmānic recension and among the first generation of the followers of Muhammad”.
Sidky’s paper relied on differences in the dotting of some words to reach this conclusion. In his AMA
https://www.reddit.com/r/AcademicQuran/comments/1cu21qs/ama_with_hythem_sidky_executive_director_of_the/l4kskgn/

he states that his conclusion can also be reached through other features besides dotting such as iʕrāb (final short vowels and Tanwin). This means that the common oral tradition included iʕrāb.


2- Early Arabic writing wasn’t phonetic.
According to the Jallad-Putten model, the Uthmānic standard text was a phonetic text written in Old Hijazi. The mismatch between the standard text and the pronunciation is the result of the later introduction of the language of poetry into the Quran.
But this can be challenged by early poetry inscriptions.
Since that Arabic poetry follows rigid meters, we know that the pronunciation of these poetry verses included Tanwin and other Classical Arabic features that don’t get spelled in the Quran. Yet Tanwin and the other features aren’t spelled in these poetry inscriptions.

The first inscription:

This inscription includes the date it was inscribed which is the year 98 AH. The location is near Mecca. The inscription contains the following poetry verse:

أفنى الجديدَ تقلّبُ الشمسِ * وطلوعُها من حيثُ لا تُمْسي
وطلوعُها بيضاءَ صافيةً * وغروبُها صفراءَ كالوَرْسِ
afna l-jadīda taqallubu sh-shamsī
wa-ṭulūʕuhā min ḥaythu lā tumsī
wa- ṭulūʕuhā bayḍāʾa ṣāfiyatan
wa-ghurūbuhā ṣafrāʾa kal-warsī

Source:

https://www.awbd.net/article.php?a=68

 

The following 3 words were non-phonetically spelled:

1- ṣāfiyatan. Spelled in the inscription as ṣāfyh صافية

The verse’s meter requires the word to be pronounced with the final short vowel ‘a’ and Tanwin: ṣāfiyatan.

But the word is spelled in accordance with the Quranic and standard Arabic spelling: The feminine ‘t’ is spelled as ‘h’ and the Tanwin isn’t spelled.
If the word was phonetically spelled, it would have looked like this:

ṣāfytn صافيتن

The verse also includes these two words:

الشمس
ash-shamsī (spelled as alshms)

كالورس

kal-warsī (spelled as kalwrs)

 

These two words employ a poetic license used in most of poetry. In Arabic grammar, when a word comes at a pausal position such as the end of a verse, the final short vowel and Tanwin are dropped. But for the sake of creating rhymes, poetry allows the retention and lengthening of the final short vowel of the last word of a verse. This poetic license is used with these two words. The final short ‘i' vowel is retained and lengthened into ‘ī’. The rhyme and the meter of the verses require that these two words end with ‘ī’ yet it’s not spelled. Which means the inscription isn’t phonetic. If it was, then it would have spelt the two words as:

الشمسي

alshmsī

كالورسي

kalwrsī

  

The second inscription:

An inscription by the son of the prophet’s poet Ḥassān Bin Thābit.

أنا الحسام ابن حسان الذي قال

الملحقين حليفهم بصريحهم والخالطين فقيرهم بالوافر

Source

 

The inscription says “I am Al-Ḥusām Bin Ḥassān who said…”. What follows is a poetry verse by his father.
The last word of this verse is bil-wāfirī. Spelled as balwāfr بالوافر

The rhyme and the meter require the retaining of the last ‘i' vowel and lengthening it into ‘ī’. But the word is spelled without a final ‘ī'. Which means the inscription isn’t phonetic. If it was then the word would have been spelled as

بالوافري

balwāfrī

 

The third inscription:

This one is discovered by the desert team. They date it to the first century AH. The poetry verse says:

أقول لما أتاني ثمَّ مصرعه

لا يبعد الرمح ذو النصلين والرجل
aqūlu lammā atānī thamma maṣraʕuhū
lā yabʕadi r-rumḥu dhu n-naṣlayni war-rajulū

Source

 

1- The word

مصرعه

maṣraʕuhū (spelled as mṣrʕh)

The ū vowel that’s attached to the ‘h’ pronoun isn’t spelled in accordance with Quranic and Classical Arabic spelling. The meter requires the pronunciation of this ū. Which means this verse isn’t written phonetically. If it was then the word would have looked like this

mṣrʕhū مصرعهو

 

2- The last word of the verse is:

والرجل

war-rajulū. Spelled as walrjl.

The rhyme and the meter require the retaining of the last ‘u' vowel and lengthening it into ‘ū’. But the word is spelled without a final ‘ū'. If the word was phonetically spelled, it would have looked like this:

والرجلو

walrjlw

The pausal spelling rule revised
These inscriptions don’t prove the validity of the pausal spelling rule proposed by later authors as an explanation for the mismatch between the Quranic text and its pronunciation. The people of the Hijaz had a spelling that reflected their Iʕrāb-less spoken register. They stuck to these spelling forms even when they wrote texts that differ in pronunciation from the spoken register, texts such as poetry.
The people of the Hijaz originally spoke in a language that had full Iʕrāb. Over time, this language lost Iʕrāb. The people of the Hijaz were also composing poetry in a language that lost Iʕrāb only at pause. After the language of poetry was applied on the Quran, later authors noticed that the Quranic spelling matches the pronunciation only at pause. This led them to assume that every word in the Quran was spelled as if the word was at pause.

 

3- The spoken register and the literary register.
According to the Jallad-Putten model, the people of the Hijaz didn’t compose poetry in the classical form of Qaṣīdah. But sometime during the Umayyad rule, the Qaṣīdah register became the high register while Old Hijazi become only a performative register.
But this can be challenged by early poetry inscriptions.  

The evidence from early Arabic texts written with Greek letters proves that the spoken language lacked a full case system. But this still doesn’t close the door on the possibility that that there was a poetry register. The Islamic tradition reports a numerous number of poems dated to the time of Muhammad. Many of which are attributed to Ḥassān Bin Thābit who had the title of “The Prophet’s poet”.  These poems are composed in the Qaṣīdah form where specific rigid meters are followed and full Iʕrāb is used. The previously mentioned inscription by Ḥassān Bin Thābit’s son proves that his father indeed composed poetry in the Qaṣīdah form:
https://x.com/OfNajd/status/1243940166538051584

The inscription says “I am Al-Ḥusām Bin Ḥassān who said…”. What follows is a poetry verse by his father. The verse is in the Qaṣīdah form.
الملحقين حليفهم بصريحهم
 والخالطين فقيرهم بالوافر
al-mulḥiqīna ḥalīfahum biṣarīḥihim (or biṣarīḥihum)
wal-khāliṭīna faqīrahum bil-wāfirī

On the same rock there’s another inscription that says:

“I am Abdul Raḥmān the son of Ḥassān Al-Ḥusām.”

The desert team’s report shows that the Islamic sources do report that Ḥassān Bin Thābit had the title of “Al-Ḥusām” and that he had a son named Abdul Raḥmān. In the first inscription, “Al-Ḥusām” is the name of Ḥassān’s son. In the second inscription, “Al-Ḥusām” is the title of Ḥassān. The desert team concludes that Ḥassān had a second son he named after his own title. The Islamic sources missed reporting on the second son.

These two inscriptions prove that Muhammad’s poet did compose poetry in the Qaṣīdah form.


Why Classical Arabic replaced the Old Hijazi reading of the Quran
Now that we have established that the people of the Hijaz had a vernacular register and a poetry register, it shouldn’t be surprising that reciting the Quran using the literary register started to appear at a very early stage, maybe even during Muhammad’s lifetime. It’s also not surprising that this new style of recitation eventually took over and replaced the original style. There are two reasons for this.
1- Early Muslims were not focused on preserving the original language of the Quran. This is proven by the fact that the reading traditions are full of features that don’t belong to the Qurayshi dialect. We even have a clear statement from Khalaf, the eponymous reader of one of the ten canonical readings, who says: “ (the tribe of) Quraysh did not pronounce the hamzah; the hamzah was not part of their language. Rather, the readers pronounced it following the languages of Arab tribes other than Quraysh.”

((وقريش لا تهمز، ليس الهمز من لغتها، وإنما همزت القراء بلغة غير قريش من العرب)). (رسم  المصحف لغانم قدوري ص357)

2- The addition of final short vowels beautifies the recitation and makes it more melodic and easier to sing.

How Final short vowels and Tanwin beautify the recitation
Applying final short vowels makes any Arabic text more melodic and easier to sing especially that these added short vowels break up clusters of consonants. For a native Arabic speaker like me, the difference in beauty between a Classical Arabic recitation and an Old Hijazi recitation is night and day. It’s true that the Classical Arabic recitation misses the hundreds of Old Hijazi internal rhymes, but the tradeoff is more than worth it. Final short vowels and Tanwin make most Quranic verses considerably more melodic, especially because they break up consonant clusters. Take for example the second verse of the first Surah of the Quran:

Old Hijazi:

al-ḥamd li-llāh rabbi l-ʕālamīn

Classical Arabic:

al-ḥamdu li-llāhi rabbi l-ʕālamīn

Note: Since that final short vowels in Old Hijazi are retained in construct, the word “rabb” takes the final short vowel ‘i’ in both renditions.

Notice how the addition of ‘u’ to “al-ḥamd” breaks up the cluster of consonants of “mdl”:
al-ḥamd(u) li-llāh(i).
This break up makes the text easier to sing.

Melodically, a long vowel at the start of a cluster, such as in āCC, īCC, or ūCC, functions similarly to a consonant. Consequently, a sequence like āCC creates the same prosodic weight as a cluster of three consonants (CCC). This comes in accordance with Arabic grammar which doesn’t have the concept of “cluster of consonants” but instead it has the concept of “the meeting of two vowelless letters” التقاء الساكنين. By taking this concept into consideration, the addition of final short vowels breaks up another cluster of consonants in the verse which is “āhr”:

al-ḥamd(u) li-llāh(i) rabbi

In the first Surah of the Quran, which consists of six verses, the addition of final short vowels breaks up six clusters of consonants. Which makes the recitation more melodic.

Here are the next 4 clusters in the Surah:
1:5

إِيَّاكَ نَعۡبُدُ وَإِيَّاكَ نَسۡتَعِينُ
Old Hijazi:
iyyāk naʕbud wa iyyāk nastaʕīn
Classical Arabic:
iyyāk(a) naʕbudu wa ʾiyyāk(a) nastaʕīn

1:7
صِرَٰطَ ٱلَّذِينَ أَنۡعَمۡتَ عَلَيۡهِمۡ غَيۡرِ ٱلۡمَغۡضُوبِ عَلَيۡهِمۡ وَلَا ٱلضَّآلِّينَ

Old Hijazi:
ṣirāṭ al-ladhīn anʕamt ʕalayhum ghayr al-maghḍūb ʕalayhum wa la ḍ-ḍāllīn
Classical Arabic:
ṣirāṭa l-ladhīna anʕamt(a) ʕalayhim ghayri l-maghḍūb(i) ʕalayhim wa la ḍ-ḍāllīn

The word “ḍ-ḍāllīn” has the cluster “āll” which remains unbroken in classical Arabic.

The addition of Tanwin also breaks up clusters of consonants since that Tanwin is always preceded by an added final short vowel. See this example which breaks up the two clusters “wml” and “ybf”:

﴿لِيَوْمٍ لَا رَيْبَ فِيهِ﴾

 3:9

Old Hijazi:
liyawm lā rayb fīh

Classical Arabic:
liyawm(in) lā rayb(a) fīh

The Ideological application of Iʕrāb
This note emphasized by Putten (Quranic Arabic p.187) explains why there’s no trace of any reading that lacks Iʕrāb. Putten points out that there are a hundred prophetic and non-prophetic narrations admonishing people not to read the Quran without Iʕrāb. He then says:
“These narrations reveal something important: Later readers and grammarians thought that any form of language could not be considered al-ʕarabiyyah unless it was supplied with ʔiʕrāb, which being the maṣdar of the C-stem of the root √ʕrb literally means “to make Arabic”. Therefore, there is no al-ʕarabiyyah without ʔiʕrāb. It is not surprising that so few of the traditions seem to acknowledge a once existing form of Quranic Arabic that did not have ʔiʕrāb even if it did exist; By the time the narrations were fabricated, and certainly when the reading traditions are canonized with Ibn Muǧāhid in the fourth Islamic century, the superiority of the ʕarabiyyah that had ʔiʕrāb/tanwīn was well-established and completely unassailable.”


The Uthmānic text vs the ancestral Uthmānic oral tradition

The previously shown evidence of ʕādallūlā and Haytham Sidqi’s work creates a contradiction: They prove that the ancestor of the canonical readings had Iʕrāb. How can this be consolidated with the strong evidence from the Uthmānic text which proves the lack of Iʕrāb? The evidence isn’t limited only to internal rhymes. Putten showed that the feminine ‘t’ is only spelled as ‘t’ in construct yet it’s never spelled as ‘t’ in non-construct which is identical to how the feminine ‘t’ is treated in modern Arabic dialects which have no Iʕrāb.
To resolve this contradiction, I propose that Iʕrāb at the time of Uthmān was only applied to melodic recitation. But in normal speech or sermons, Iʕrāb wasn’t applied on the Quran. This means that the Quran was dictated to the Uthmanic scribes without Iʕrāb, which explains the bits of evidence found by Putten such as the treatment of the feminine ‘t’.
This distinction between melodic recitation and normal speech is inspired by how Muslims treat the Quran today: Tajweed rules are applied to recitation, but are dropped when verses are quoted in ordinary speech or sermons.

The adoption of Hamzah sheds a light on iʕrāb
Both of the Hamzah and Iʕrāb are foreign to the original language of the Quran. Since that Iʕrāb was introduced at a very early stage, we don’t have any reports concerning its adoption. In contrast, hamzah was introduced later, leaving behind many reports on how the Hamzah was treated by early Muslims. Based on these reports it’s possible to assume that the application of Iʕrab went through the same stages of the application of Hamzah: First it was a preference, then it became an obligatory part of recitation.
The following reports show that Quranic pronunciation could vary according to performance context:
Although the Hamzah is used in all the canonical readings, there are reports showing that the Hamzah wasn’t applied in public prayers even by readers who have the Hamzah in their own readings.
The following statement is reported from two canonical readers: Ḥamzah and his student Al-Kisāʾī:
“Leaving off the hamzah in the prayer niches (miḥrābs) is a mark of mastery”. (The miḥrāb is the designated niche within a mosque where the imam stands to lead public prayers.)
«معرفة القراء الكبار على الطبقات والأعصار» (ص69):

«وقال أبو عمر الدوري: قال حمزة ترك الهمز في المحاريب من الأستاذية.»

«معرفة القراء الكبار على الطبقات والأعصار» (ص74):

«وقال أحمد بن فرح: حدثنا الدوري سمعت الكسائي يقول: من علامة الأستاذية ترك الهمز في المحاريب.»

It’s also reported that Abū ʕamr Al-Dūrī, one of the canonical readers, used to drop the Hamzah in public prayers although his own reading uses the Hamzah:
“Abū ʕamr used not to pronounce the hamzah during prayer, so I asked him about that, and he said: 'I acquired (Quranic) reading from Mujāhid, (who acquired it) from Ibn ʕAbbās, (who acquired it) from Ubayy bin Kaʕb, (who acquired it) from the Prophet.’ “

كتاب لغة قريش لمختار الغوث ص63

كان أبو عمرو لا يهمز في الصلاة، فسألته عن ذلك فقال: أخذت القراءة عن مجاهد عن ابن عباس عن أبي بن كعب عن النبي.

Ḥamzah the canonical reader also says that pronouncing the Hamzah is a mere exercise for the learner:
Ḥamzah was asked about some features of his reading: "what is this use of the Hamzah, elongation and severe cutting?" Ḥamzah replied: "This is an exercise for the learner."

«جامع البيان في القراءات السبع» (2/ 482):

«و] وقف الثوري على حمزة فقال: يا أبا عمارة، ما هذا الهمز والمدّ والقطع الشديد؟ فقال: يا أبا عبد الله، هذه رياضة للمتعلّم. قال: صدقت.»

Ḥamzah also said: “Pronouncing the Hamzah is merely an exercise. Once a man masters it, he should drop it.”

«معرفة القراء الكبار على الطبقات والأعصار» (ص70):

«وعن حمزة قال: إنما الهمز رياضة، فإذا حسنها الرجل سهلها.»

There’s also this report which shows that the Hamzah wasn’t expected to be pronounced in verses quoted in normal speech:
Abū Bakr, one of the two canonical transmitters of the canonical reading of ʕāṣim, said: “I visited ʕāṣim while he was ill, and he fainted. Then he recovered and read (the verse) : 'Then they are returned to Allah, their true Protector. Surely, His is the judgment, and He is the swiftest of reckoners' [Quran 6:62], and pronounced the hamzah. So I knew that (Quranic) reading was second nature to him."

«معرفة القراء الكبار على الطبقات والأعصار» (ص53):

«وقال أبو هاشم الرفاعي: حدثنا يحيى حدثنا أبو بكر قال: دخلت على عاصم فأغمي

عليه فأفاق، ثم قرأ: {ثُمَّ رُدُّوا إِلَى اللَّهِ مَوْلَاهُمُ الْحَقِّ أَلا لَهُ الْحُكْمُ وَهُوَ أَسْرَعُ الْحَاسِبِينَ} [الأنعام: 62] . فهمز فعلمت أن القراءة منه سجية.»

The verse that ʕāṣim read has the hamzah in three words:
إلى، ألا، أسرع
These three Hamzahs are pronounced in all canonical readings. Yet Abū Bakr was surprised that ʕāṣim pronounced these Hamzahs while he was in such a condition. So he concluded that the Quranic reading craft was a second nature to ʕāṣim.

Hamzah and the disappearance of the Old Hijazi reading
The lack of any trace of the original Old Hijazi reading shouldn’t be surprising given how the Hamzah eventually prevailed leaving no trace of any reading that lacks the Hamzah except for one: the reading of Abū Jaʕfar in the transmission of Al-ʕumarī (a non-canonical transmitter of the reading).
«النشر في القراءات العشر» (1/ 428):

«وَكَأَبِي جَعْفَرٍ مِنْ أَكْثَرِ رِوَايَاتِهِ وَلَا سِيَّمَا رِوَايَةُ الْعُمَرِيِّ، عَنْ أَصْحَابِهِ، عَنْهُ، فَإِنَّهُ لَمْ يَكَدْ يُحَقِّقُ هَمْزَةً وَصْلًا،»

It’s astonishing that there’s only one reading that entirely lacks the Hamzah given that the case for the survival of Hamzah-less readings is stronger than the case for the survival of Iʕrāb-less readings. That’s for the following reasons:
1- Unlike Iʕrāb, the application of Hamzah didn’t take an ideological turn.
2- The introduction of Iʕrāb doesn’t violate the Uthmanic text since that the people of the Hijaz didn’t spell Tanwin and other features when they wrote poetry verses in the Qaṣīdah format (as shown earlier). But the introduction of Hamzah violates the Uthmanic text in hundreds of cases where the Hamzah was either spelled as ى  or و, Such as the word ملاىكه where the letter ى   represents a ‘y’ sound: malāyikah. Yet all reading traditions pronounce this word with a Hamzah: malāʾikah. This pronunciation violates the written text since that the symbol for the hamzah in early Arabic writing was the Alef. The ى and و  where used in the Uthmanic text because the dropping of Hamzah results in the creation of the sounds y, ī, w or ū.
A reading that entirely lacks the Hamzah is a reading that matches the written standard text. So it’s surprising that only one surviving reading achieves this matching.
3- While we know from the evidence previously shown that Iʕrāb was already the standard recitation at the time of the Uthmanic recension, the Hamzah was introduced in a later stage. A report says that the people of Medina didn’t pronounce the Hamzah until Ibn Jundub (d. 106 AH) pronounced it, after which they started to pronounce 'mustahzūn' and 'istuhzī' with a hamza.

(( كَانَ أهل الْمَدِينَة لَا يهمزون حَتَّى همز ابْن جُنْدُب فهمزوا مستهزون واستهزي)). (السبعة في القراءات ص60)

According to this report, the Hamzah was first introduced into the Quranic recitation in late first or early second century AH.