Madd as Orthoepy Rather Than Orthography


 This article explores the function and distribution of the maddah sign throughout the history of the Islamic manuscript tradition. It demonstrates that, to date, descriptions have not adequately described its use, and it shows that rather than being a part of Classical Arabic orthography, medieval sources clearly indicate that the maddah sign was specifically used to express an orthoepic feature of Classical Arabic prose, namely madd, the pronunciation of vowels as overlong.


Introduction1
The maddah sign in modern Arabic orthography has only one function: it distinguishes a hamzah followed by an ā from a hamzah followed by short a, as in ‫آ‬ ‫ت‬ ‫ى‬ ʾātā 'to give' versus ‫أ‬ ‫ت‬ ‫ى‬ ʾatā 'to come' . However, when we look at medieval Arabic manuscripts, we find a much more diverse use of maddah, which frequently occurs throughout Islamic manuscript history. It occurs in places where it cannot represent the sequence hamzah followed by long ā.
In modern works on Classical Arabic grammar and orthography, if the broader use of the maddah in pre-modern manuscripts is mentioned at all, it is explained as a purely orthographic practice.
When êlif with hèmza and a simple vowel or tènwīn ‫َأ(‬ , ‫ِإ‬ , etc.) is preceded by an élif of prolongation ( ‫ـ‬ َ ‫ا‬ ), then a mere hèmza is written instead of the former, and the sign of prolongation, ‫ـ‬ ٓ mèdda or maṭṭa ( Further miniature letters were introduced: a […] small mīm-dāl derived from madd 'extension' for the word-initial ʾā or the word-final āʾ.4 Indeed, when examining vocalised Classical Arabic manuscripts, one frequently encounters this practice of the maddah sign before ʾalif followed by hamzah. It is often used instead of the hamzah sign, the maddah being sufficient to denote its presence. This, however, is neither a particularly archaic practice, nor a purely Quranic peculiarity, as Fischer suggests, but rather occurs in classical manuscripts throughout history until the modern period. Moreover, the descriptions given by these works are incomplete. While it is true that the most common context in which the maddah sign appears is above a word-final ā followed by a hamzah and not infrequently with word-initial hamzah followed by a long vowel, as Gruendler suggests, this is by no means the only context in which it appears. It can also occur above ī and ū followed by a hamzah. Moreover, it may arise just as readily word-medially as it occurs word-finally. Lastly, its spelling is triggered not only by a subsequent hamzah but likewise by a succeeding long consonant. While these uses of the maddah sign occur frequently in the Islamic manuscript tradition, the only accurate description of the spelling of the maddah that I am aware of is by Nico van den Boogert in his study of the Maghrebi script.5 It is noteworthy that in Maghrebi writing, the modern use of the maddah, to denote a hamzah followed by ā, is never used. Instead, a baseline hamzah is simply placed before the ʾalif.6 This practice has also been adopted in modern print editions of the Quran, where, for example, The following overview is a selection of dated manuscripts throughout the centuries that reveal the practice of using the maddah sign. No specific attempt has been made to be comprehensive-an impossible task considering how commonly the maddah sign is deployed-and my selection of these manuscripts is based on their availability in digitized form and whether, during my research, I had worked with them for other topics or was easily able to identify examples of the rarer contexts in which the maddah sign is used. This overview will establish that the spelling of maddah may occur above any long vowel, and that it occurs in situations of a following hamzah and a succeeding šaddah. This broad view will clearly establish that this is not archaic, but rather a standard and widespread spelling practice throughout the Islamic manuscript tradition. It is hoped that the overview will promote wider awareness and a more accurate description of this spelling practice.

Ibn al-Nadīm's Fihrist (End of the 10th Century) ms Dublin, Chester Beatty Library, Ar 3315
This copy of Ibn al-Nadīm's Fihrist (Chester Beatty Library, Ar 3315)7 was copied from the (now lost) original. It is estimated to date from between around 377 ah/987 ce (the year Ibn al-Nadīm finished his work) and 1000 ce and contains many examples of the maddah sign of ā before a hamzah (nb. not exclusively for word-final āʾ sequences), e.g.:

Quranic Madd
Those familiar with Quranic recitation will be struck by the environment in which the maddah occurs in these Classical Arabic manuscripts, as it coincides perfectly with the places where Quranic recitation requires the application of the madd, i.e. the overlong articulation of long vowels. Manuals on Quranic reading traditions typically feature a chapter containing considerable discussion of the conditions in which madd occurs and the relative lengths of the overlong vowels among the readers, and some of their specificities. We will summarize the rules as they pertain to the canonical seven readers briefly here, based on al-Dānī's Ǧāmiʿ al-Bayān: 1. When ū, ī, or ā precedes a hamzah within a word, all readers of the Quran agree that it should be read overlong, as in, for example, ʾulāāʾika, qāāʾimīn, hanīīʾan, marīīʾan, bi-sūūʾin, and ʾan tabūūʾa.11 2. When the sequence of a long vowel plus a hamzah occurs across word boundaries, there is disagreement on whether to lengthen the vowel. Most readers treat this environment in the same way, thus reading yāā ʾayyuhā, yā banīī ʾādam, and ǧāāʾūū ʾabā-hum. However, Ibn Kaṯīr, Qālūn ʿan Nāfiʿ, Warš ʿan Nāfiʿ (in the transmissions of Yūnus and al-ʾAṣbahānī), and ʾAbū ʿAmr do not lengthen the vowel in between words.12 3. All readers are also in agreement that a vowel should have an overlong pronunciation if it is followed by a long consonant, e.g. wa-lā ḍ-ḍāāllīn.13 Warš ʿan Nāfiʿ adds two more environments in which madd is applied. Firstly, he adds it to long vowels when they are preceded by hamzah, e.g. ʾūūtiya, li-l-ʾīīmāni.14 This appears to correspond to the maddah that denotes the word-initial ʾā, as it is still used in Modern Standard Arabic today, and not infrequently (albeit less often than when a hamzah follows the long vowel) in earlier manuscripts. Finally, it has been transmitted that Warš ʿan Nāfiʿ applied madd to ay and aw, e.g. šayyyyʾin, ka-hayyyyʾah, and as-sawwwwʾi.15 In later, vocalised Quranic manuscripts, such cases of madd are, understandably, written with the same maddah that we see in these non-Quranic manuscripts. To a large extent, the places in Quranic manuscripts where the maddah is written corresponds to where the maddah is also written in non-Quranic Classical Arabic texts, as examined above.
The use of madd across word-boundaries seems to be typical for Quranic recitation and does not find a clear reflex in the non-Quranic manuscript tradition.

7
Madd amongst the Grammarians The clear parallels between the places where the maddah sign is written in Classical Arabic manuscripts and the environment in which madd is applied in Quranic recitation suggest that this cannot be coincidence. Thus, while the use of madd is usually considered a feature that is specific to Quranic recitation,16 the distribution strongly suggests that the practice of making these vowels overlong in these environments was not unique to Quranic recitation, but rather part of general Classical Arabic orthoepy. As one would expect, the topic is also discussed by grammarians, explicitly not in the context of describing the proper recitation of the Quran, but simply with respect to the proper pronunciation of Classical Arabic prose. To my knowledge, the earliest grammarian to discuss this phenomenon explicitly is Ibn Ǧinnī (322-411ah/941-1002 ad) in his influential grammatical work al-Ḫaṣāʾiṣ in bāb fī maṭl al-ḥurūf "the chapter on the stretching of the letters".17 He first discusses the definition of the long vowels, citing the examples qām 'he stands' , sīr bi-h 'travel with him!' , ḥūt 'whale' , kūz 'small jug of clay ' ,14 Ibid.,193. 15 Ibid., 202f. 16 E.g. Kristina Nelson, "Tajwīd," -Naǧǧār (Cairo: al-Maktabah al-ʿIlmiyyah, 1952), iii, 124ff. in the time of Sībawayh (d. 180 ah/796 ad), who in bāb al-maqṣūr wa-l-mamdūd 'the chapter of the shortened and the lengthened'18 says: The shortened (using manqūṣ instead of maqṣūr) are each word among the sisters of the yāʾ and wāw (i.e. roots with yāʾ or wāw as a final root consonant) where its yāʾ or wāw occurs after a letter that carries a fatḥah, so its shortening is that they are replaced by an ʾalif in the place of the yāʾ and wāw and it does not end in the accusative, nominative or genitive.
Turning to the lengthened forms he says: As for what is lengthened, it is everything that occurs when a yāʾ or a wāw would come after an ʾalif. So things like this are known to be lengthened, and this is for example with al-istisqāāʾ 'praying for rain' because istasqaytu 'I prayed for rain' is (of the form) istafʿaltu, just like istaḫraǧtu 'I moved out' , so if you want the maṣdar you know that it is certain that there is supposed to be a yāʾ after the ʾalif just like it is certain for the ǧīm (of istaḫraǧtu) to be present in the maṣdar (istiḫrāǧ) after the ʾalif.
Other examples he gives include: al-ištirāāʾ 'purchase'; al-ʾiʿṭāāʾ 'granting'; al-ʿuwāāʾ 'howling'; and ad-duʿāāʾ 'prayer' , as well as several nouns that are only recognised by making them plural, such as as-samāāʾ 'the sky' , as then the semivowel that is replaced by the hamzah returns, as in as-samāwāt 'the heavens' . Interestingly, in the discussion of these nouns with a lengthened final vowel, at no point does Sībawayh point out that the root consonant wāw or yāʾ is replaced by a hamzah. To Sībawayh, it seems that the most salient feature of the ʾalif mamdūdah was indeed what the name suggests, i.e. that is was lengthened.19 This suggests that already in Sībawayh's time the orthoepic pronunciation of an overlong ā (and likely also ī and ū) when it was followed by a hamzah was considered normative. 18 ʾAbū Bišr ʿUṯmān Sībawayh, Kitāb Sībawayh, ed. ʿAbd al-Salām Muḥammad Hārūn (Cairo: Maktabat al-Ḫānǧī, 1988), iii, 536ff. 19 For a similar case of the use of madd versus qaṣr to describe ʾalif followed by a hamzah, see Sībawayh's comment ʾa-lā tarā-hum qālū: Zakariyyāʾūna fī-man madda, wa-qālū zakariyyawna fīman qaṣara "Do you not see that they say Zakariyyāʾūna (for plural of Zaka-riyyāʾ) for he who lengthens and they say zakariyyawna (for the plural Zakariyyā) for those that shorten" (Ibid., iii, 394).

Conclusion
Contrary to common description, the maddah sign is not primarily used to write syllable-initial ʾā in Classical Arabic orthography. Rather, it frequently occurs in environments where ā, ī, or ū precede a hamzah or a long consonant. This is exactly the environment where, even today, in Quranic recitation, one is to pronounce these vowels overlong. Ibn Ǧinnī's description makes it clear that, in his time, this type of orthoepic overlong vowel realization (madd) was not considered an exclusively Quranic practice. Instead, it seems to have been part of the historical orthoepic pronunciation of any form of Classical Arabic prose. In light of this, we should conclude that the use of the maddah sign was not so much an orthographic practice deployed to indicate that a hamzah follows or precedes a long vowel, as has been suggested by descriptions at the start of this article, but rather it should be seen as a sign that specifically denoted a difference in pronunciation, namely that of an overlong vowel.