not one, but three, tens, hundreds

Most Indians, most Arabs, most Turks, think that all editions of the qurʾān are the same.
And they are right:
For 150 years most copies printed in Kara­tchi, Delhi, Dhakka, Johannis­burg are the same.
Since 1985 most printed east of Libya and west of Persia follow the ortho­graphy of the 1952 Egyptian state edition,
and most copies printed in Turkey (since 1950 ???) are practically identical.
Nevertheless, Gabriel Said Rey­nolds is completely wrong, when he states

the various editions of the Qur’an printed today (with only extra-ordinary excep­tions) are identical, word for word, letter for letter.
"Introduction" to The Qur'ān in its Histori­cal Context, Abingdon: Routledge 2008, p. 1

from left to right: Syrien, Qaṭar, Kuwait (al-Ḥaddād), Bahrain, Saudia, VAE (both UT1), Dubai, Saudia(UT2), Kuwait (UT1), Oman, Kerbala, Ägypten (Abu Qamar)

Yes, nowaday most maṣāḥif produced in the Arab mašriq are similar, but Morocco, Libya, Sudan, Turkey, Tarta­ris­tan, Brunai, Indo­nesia follow different rules, and the Indian Stan­dard (Pakistan, Bangla Desh, UK, South Africa, Surinam, Nepal, Ceylon) is numeri­cally more important and quite diffe­rent. "Nowadays" because before 1980 a Ottoman muṣḥaf written by Ḥafiz ʿUṭmān the Elder (1642‒1698) was pre­valent in Syria, and two Ottoman maṣāḥif written by Ḥasan Riḍā and Muḥammad ʾAmīn ar-Rušdī res­pectively were pro­duced for Dīwān al-Awqāf al-ʿIrāqī (still 1980 the govern­ments of Qaṭar and Saʿūdī ʿArabia had copies printed of the one based on Rušdī ‒ and 1415/1994 in Tehran): It took some seventy years before the 1924 edition (or rather its 1952 offspring) had created a regional stan­dard.

Because there are THREE well established standards and a few in Indonesia, a new one in Brunei, several (competing ones) in Iran and many all over Africa ‒ where we do not only find different ways of writing the same reading (Ḥafṣ ʾan ʿĀṣim) but three more trans­mis­sions (Warš, Qālūn, ad-Dūrī ʿan Abī ʿAmr). And 100 years ago, maṣāḥif were less stan­dardized.
There are many more printed in Damas­cus (or Bairūt because of the war), pro­duced in ʿAmman and the UAE and pub­lished on the world wide web, but these are mainly for study, not for devotion.
But here I will not focus on the readings (and their trans­mis­sions), but on different ortho­graphies (of the trans­mission Ḥafṣ ʿan ʿĀṣim).
Already 35 years ago Adrian Alan Brockett found out that the 1342/1924 King-Fuʾād-Edi­tion had not estab­lished THE stan­dard, that even the suc­ces­sor of al-Ḥusainī al-Ḥad­dād as the chief reci­ter of Egypt ‒ hence main editor of the "second edition" of 1952 ‒ ʿAlī Muḥammad aḍ-Ḍabbāʿ (1304/1886-1380/1960), had edited different editions and approved of yet more dis­similar ones.

Brockett studied editions at a time when only Ḥafṣ and Warš were prin­ted.
Today one finds many edi­tions of Qālūn, some of Dūrī and both prin­ted ones and just pdfs for most of the others, plus many edi­tions about the 20 cano­nical trans­missions, plus sound files of reci­ta­tions of most trans­missions.       When Brockett wrote, the King Fahd Complex had not started to pub­lish dif­ferent vari­ants ‒ Ḥafṣ, Šuʿba, Warš, Qālūn, Dūrī, as-Sūsī writ­ten by ʿUṭmān Ṭāhā plus an Indian Ḥafṣ ‒ but he had no­ti­ced that Gulf States pub­lished a) in the new Egyp­tian style, b) an Ottoman muṣ­ḥaf (the muṣḥaf of M­uḥam­mad ʾAmīn ar-Rušdī with minor mo­di­fi­ca­ti­ons), c) in the Indian style.

1952:

Brockett's thesis is still the best English "book" avail­able on dif­ferences bet­ween copies of the qurʾān, al­though it was researched before the inter­net faci­li­tated research, before Uni­code made it easier to re­pro­duce Arabic script,
before it was easy to get hold of all the cano­nical trans­mis­sions and most of the thousands of variant readings (col­lected in three dif­ferent editions).
His main con­clu­sions ‒ the oral trans­mission and the one in writing re­in­forced each other, con­trolled each other, never were left with­out the other,
and there is no single standard of writing, and no single stan­dard of reciting the qurʾān,
and the dif­feren­ces between trans­mis­sions (and within trans­missions) are minor, they never change the meaning of a para­graph ‒
stand intact. But it was a thesis, no published book.
Because the young student was not allowed to have it read by fellow researchers, it is full of mistakes,
mis­takes which would have been elimi­nated before pub­lication as a book.
I personally have no use for Brockett's "trans­literation", which is neither that nor a tran­scription.
I am sure that Brockett ‒ as many readers ‒ did not know what the two terms mean:
a trans­literation must render the Arabic letters faith­fully and must be rever­sible (not neces­sarily pro­nounce­able),
a transcription must render the sound of the words faith­fully, must be pro­noun­ce­able, should be read­able after some instruc­tion, but has not to be reversable, because different sequences of letters can be pro­nouned (hence transcribed) the same way.
I personally, hate his termino­logy, but at least he defines his ‒ odd ‒ terms at the outset:
"graphic" means: part of the rasm,
"oral" means: not part of the rasm.
I say: utter nonsense!
Both the rasm and the later signs (dots, hamza, waṣl, shadda, fatḥa, kasra, ḍamma, signs for imāla, tasḥīl, išmām etc.) are gra­phical,
and have to be pro­nounced = are oral ‒ but there are some otiose letters, which have to be written in a real trans­literation (as in R-G Puin's).
"oral only" is closer to what he means, but "in the oldest manu­scripts not written, at that time: only recited" is it.
Sorry, "oral" is not good enough.
I hardly can read his "trans­cription". Why does "a wavy line" means some­times "oral", sometimes "lengthened"?
Anyhow, here and now, there is no need for Brockett's "trans­litera­tion", we have Arabic letters!
In spite of my criticism, his thesis is a great work of scholar­ship ‒ and tre­men­dous work, done before we just googled dif­ferent edi­tions of the qurʾān.
The content of this blog and my German one, you can find as book.

Gizeh 1924 <> Cairo 1952 and after

Page 775 of the Amiriyya print (page 574 in editions that end on page 604) is remarkable because in the first line allan is sometimes written ان لن sometimes الن . There is no difference in meaning, no difference in pronounciation.
But it is important to some: they deliberately "correct" the spelling. Here now, two pages from the Amiriyya, both with الن




There are three differences on this page between the 1924 and 1952 edition, typical differences found throughout the muṣḥaf -- there are more than 800 of these -- plus four minor corrections.

To show that the changes did not stop 1952, I have copied two version distributed by the King Fahd Complex into the Amiriya-frame:
first ʿUṭmān Ṭāha 1

then ʿUṭmān Ṭāha 2

On the next pair there is no sura end, no sura title, but again one changed pause sign and on the very last word the hamza has moved from above to below the line (which is one of the four corrections mentioned in the afterwork to "the second printing").
-- the second page is not from the Amiriya but from a Bairut print, hence the page number is on top of the page and the catch word is missing.



On the last pair there is only one difference: kalimatu (line 5) is written with ta maftuḥa vs. marbuṭa.

Giza 1342/3 1924/5

The Giza print
‒ is not an Azhar Quran
‒ did not trigger a wave of Quran printings
    because there was finally a fixed, authorised text.
‒ did not immediately become the Qur'an accepted by both Sunnis and Shiites

‒ did not contribute significantly to the spread of Ḥafṣ reading;
    it was not published in 1923 or on 10/7/1924.
But it drove the grotty Flügel edition out of German study rooms,
‒ had an epilogue by named editors,
‒ gave his sources in it,
‒ adopted ‒ except for the Kufic counting,
    and the pause signs, which were based on Eastern sources.
    ‒ the Maghrebi rasm (largely after Abū Dāʾūd Ibn Naġāḥ)
    ‒ the Maghrebi small substitute vowels for elongation
    ‒ the Maghrebi subdivisions of the thirtieths (but without the eighth-ḥizb)
    ‒ the Maghrebian baseline hamzae before Alif at the beginning of the word (ءادم instead of اٰدم).
    ‒ the Maghrebic spelling at the end of the sura, which assumes that the next sura is spoken immediately afterwards (and without basmala): tanwin is modified accordingly.
    ‒ the Maghrebic distinction into three kinds of tanwin (one above the other, one after the other, with mīm)
    ‒ the Maghrebic absence of nūn quṭni.
    ‒ the Maghrebic non-spelling of the vowel shortening.
    ‒ the Maghrebic (wrong) spelling of ʾallāh.
    ‒ the Maghrebī (and Indian) attraction of the hamza sign by kasra

in G24 the hamza is below the baseline ‒ in the Ottoman Empire (include Egypt) and Iran the hamza stays above the line


















A new feature was the differentiation of the Maghrebic sukūn into three signs:
    ‒ the ǧazm in the form of an ǧīms without a tail and without a dot for vowel-lessness,
    ‒ the circle for never to be pronounced,
    ‒ the (oval) zero for "only pronounced if paused".
    ‒ the absence of any sign to signify "not to be pronounced".
Further, word spacing,
baseline orientation and
exact placement of dots and dashes.


Nor was it the first "inner-Muslim Koran print".
Neuwirth may know a lot about the Koran, but she has no idea about Koran prints,
because since 1830 there have been many, many Koran prints by Muslims.
and Muslims were already heavily involved in the six St. Petersburg prints of 1787-98.
It was not a type print either, but ‒ like all except Venice, Hamburg, Padua, Leipzig, St.Petersburg, Kazan and the earliest Calcutta ‒ planographic printing, albeit no longer with a stone plate but a metal plate.

Nor was it the first to claim to reproduce "the rasm al-ʿUṯmānī".
Two title pages of Lucknow prints from 1870 and 1877.

In 1895, a Qur'an appeared in Būlāq in ʿuṯmānī rasm, which perhaps meant "unvocalised". Kitāb Tāj at-tafāsīr li-kalām al-malik al-kabīr taʼlīf Muḥammad ʿUṯmān ibn as-Saiyid Muḥammad Abī Bakr ibn as-Saiyid ʻAbdAllāh al-Mīrġanī al-Maḥǧūb al-Makkī. Wa-bi-hāmišihi al-Qurʼān al-Maǧīd marsūman bi'r-rasm al-ʿUṯmānī.
Except for the sequence IsoHamza+Alif, which was adopted from the Maghreb in 1890 and 1924 (alif+madda was not possible, since madda was already taken for elongation), everything here is already as it was in 1924.

Incidentally, the text of the KFA is not a reconstruction, which Bergsträßer al-Ḥusainī al-Ḥaddād simply believed: He does not follow Abū Dāʾūd Sulaiman Ibn Naǧāḥ al-Andalusī (d. 496/1103) exactly, nor Abu ʿAbdallah Muḥammad ibn Muḥammad al-Ḫarrāz (d. 718/1318), but (except in about 100 places) the common Warš editions.
Also, the adoption of many Moroccan peculiarities (see above), some of which were revised in 1952, plus the dropping of Asian characters ‒ plus the fact that the epilogue is silent on both ‒ is a clear sign that al-Ḥusainī al-Ḥaddād al-Mālikī adapted a Warš edition.
All Egyptian readers knew the Warš and Qālun readings. As a Malikī, al-Ḥusainī al-Ḥaddād ‒ not to be confused with the scribe Muḥammad Saʿd Ibrāhīm al-Ḥaddād ‒ probably knew Warš editions even better than most.
There was the text, sup­posedly establi­shed in 1924, not only in the Maghreb and in Cairene Warš prints, but also already set in Būlāq in the century before.

Now to the date of publication.
One finds 1919, 1923, 1924 and 1926 in libraries and among scholars.
According to today's library rules, 1924 is valid, because that is what is written in the first printing.

But it is not true. It says in the work itself that its printing was completed on 10.7.1924. But that can only mean that the printing of the Qurʾānic text was completed on that day. The dedication to the king, the message about the completion of the printing, can only have been set afterwards; it and the entire epilogue were only printed afterwards, and the work ‒ without a title page, without a prayer at the end ‒ was only bound afterwards ‒ probably again in Būlāq, where it had already been set and mounted ‒ and that was only in 1925, unless ten copies were first bound and then "published", which is not likely.

India 1800 Long vowels

Gabriel Said Reynolds and others say that all Qur'anic texts are identical: letter for letter.

the various editions of the Qur'an printed today (with only extra-ordinary exceptions) are identical, word for word, letter for letter.
"Introduction to The Qur'an in its Historical Context, Abingdon: Routledge 2008, p.1.

Nonsense! There are probably a thousand different ways of writing or typesetting Qurans.
That does not mean that the prints say different things. They don't. They are similar enough -> mean the same. The differences that the exact same text allows in inter­pretation are certainly 100 times more signi­ficant than all the differences between different prints. Many differences are purely orthographic (such as folx­heršaft and Volks­herrschaft, night and nite, le roi and le rwa), others change the sense of a word, even a sentence, but do not really change the passage.
I am not at all concerned with contradictions in the Qur'an, with differences in content between one and another, I am only concerned with differences in ortho­graphy (that is, the spelling rules and particular cases).
Nor am I concerned with the differences between the seven/ten canonical readers, the four­teen/twenty transmitters, the hundreds of tradents. These primarily concern the phonetic structure (sometimes a "min" or "wa", an alif or a consonant doubling more or less); the variants only say whether a vowel is lengthened fivefold or threefold, whether the basmala is repeated between two suras or a takbir is spoken before a particular one. I am not concerned with all this.
I am interested in the differences between Ottoman and Moroccan, Persian and Indian Korans ‒ and how the official Egyptian Koran of 1924 differs from those before it. Because there is a lot of nonsense circu­lating about this.
Qurans differ in a hundred ways. I will not present this systema­tically. For example, reading style, writing style, lines per page, whether verses may be spread over two pages, whether 30th must begin on a new page, whether rukuʿat are displayed in the text and on the margin, whether verses have numbers and whether pages have custo­dians, whether there are one, three, four, five, six ... or sixteen pause signs. All this can occur, but will not be discussed.
I focus attention on two points:
the spelling of words, the Koranic vocabulary, so to speak ‒ although (unlike in the Duden) the same word is not to be written the same way in all places;
the rules of how vowel length, shortening and diph­tongs are notated, like assimi­lation of consonants. I am particularly interested in prints.
There are two main spellings/set of rules: African (Maghrebi, Anadlusian, Arabic) and Asian (Indo-Pakistani, Indonesian, Persian, Ottoman): Africans always need two signs for long vowels: a vowel sign and a matching elongating vowel letter; if the latter is not in the rasm, it is added in small (or a non-matching one is made suitable by a Changing-Alif).
Asians have three short vowel signs and three long vowel signs (plus Sukūn/Ǧazm). But according to today's IPak rules, for ū and ī, one uses the short vowel signs IF the matching vowel letter follows (which gets a ǧazm). With long ā, Persians and Otto­mans/Turks always used the long vowel sign; Indians today use it only if no alif follows (i.e. wau, [dot­less] yāʾ or no vowel at all); if an alif follows, the consonant before it only gets a Fatḥa. In the case of long-ī, Persians and Ottomans always used the Lang-ī sign (regard­less of whether it is followed by yāʾ or not); Indians today proceed simil­arly to ā: if it is not followed by a yāʾ, the long-ī sign is used: before yāʾ, however, there is (only) Kasra and the yāʾ gets a ǧazm. (According to IPak, sign-less letters are silent!).
For long ū, Ottomans put "madd" under a wau; for the elongated personal pronoun -hū, the elongation remains unnotated. Indians and Indonesians use the long ū sign but the short u sign before wau, while before 1800, Indians always used the long-ū-sign, following wau remained without any sign was thus silent (to be ignored when reading) ‒ if it is second part of the diphtong au, it got and gets a Ǧazm, thus is to be spoken. Always the long ī sign. Always the long-ā-sign. In other words:

In 1800, there were two systems of noting long vowels: the Maghre­b­ian, which always included two parts, a vowel sign (fatḥa, kasra, ḍamma, imāla-point) and a lengthening vowel (belonging to the rasm or a small comple­ment). And an Indian system based entirely on long vowel signs, in which the vowel letters present in the rasm were com­pletely ignored. The Maghrebi system is used today in Africa and Arabia. The Indian system is used in weakened forms in Turkey, Persia, India and Indone­sia. In India and Indone­sia, IPak applies, where long ā continues to be used before (dotless) yāʾ, but before alif it has been replaced by fatḥa (like in the African system) Before ī-yāʾ / ū-waw stand kasra / ḍamma; abobve the vowel letter stands ǧazm ‒ otherwise they had no influence on pronounciation. The old Indian system only applies where no vowel letter follows. How widespread this clear Indian system was, I do not know. I came across several manuscripts using it, but no print.

Impressum

Angaben nach §5 TMG:
Arno Schmitt
Gustav-Müller-Str. 10
10829 Berlin

Kontakt:
arnoas@live.de

Verantwortlich für den Inhalt nach §55(2) RStV:
Arno Schmitt

The Ḥafṣ ʿan ʿĀṣim King Fuʾād 1924/5 edition

I have just published an essay on Qurʾān prints on Amazon:
I want to blog about that and from it here.
(this is my German post translated by deepl)
In the course of time I will probably bring everything from the book - but slowly...
Since 1972, when thousands of very old Qurʾān frag­ments were discovered in a walled-up attic of the Great Mosque of Ṣanʿāʾ, more precisely since 2004, when Sergio Noga Noseda was allowed to pro­duce high-resolution colour photo­graphs, since scholars have recognised that leaves kept in up to seven different collec­tions formed one codex and that they can be studied thanks to online and printed publications.
Since thousands of short texts carved in stone from Syria, Jordan and Sa'udi Arabia can be read (ever better), research into the Arabic language and script of the cen­turies immedia­tely before and after Muḥammad has been the most exciting part of Islamic studies.
Since the destruction of the Twin Towers in Man­hattan, reflections on Islam as a late ancient civili­sation and/or religion related to Judaism and Christia­nity have been particularly popular.
Unfortunately, experts in these inter­esting fields also comment on a subject they have not studied ‒ because it is not inter­esting enough - and write almost nothing but nonsense about it.

The field of printed editions of the Qur'an needs to be cleaned up. And that is what I want to do here. Many German Orienta­lists refer to the official Egyp­tian edition of 1924/5 as "the stan­dard Qur'an", others call it "Azhar Qur'an". Some call it "THE Cairo Edition/CE" ‒ utter nonsense. Many false ideas circulate about the King Fuʾād edition, the Giza Qur'an, the Egyptian Survey Authority print (المصحف الشريف لطبعة مصلحة المساحة المصرية), the 12-liner (مصحف 12 سطر). Some believe they are looking at a manu­script, Andreas Ismail Mohr and Prof. Dr. Murks call it "type printing". Yet the epilogue ‒ from 1926 even more clearly than the first one (1924/5) ‒ makes every­thing clear: The book written by Egypt's šaiḫ al-maqāriʾ Muḥammad ibn ʿAlī ibn Ḫalaf al-Ḥusainī al-Mālikī aṣ-Ṣaʿīdī al-Ḥaddād (1282/1865-1357/ 22.1. 1939) ‒ not to be confused with the calli­grapher Muḥammad ibn Saʿd ibn Ibrāhīm al-Ḥaddād (1919-2011) ‒ was set in Būlāq with five tiers per line (pause signs; fatḥa, damma, sukūn; letters [for baseline hamza includ­ing the vowel sign]; kasra; spacing). ((added later: If you want to see/understand what was made "between Būlāq and Giza"/between type setting and printing" have a look at the Hyderabad print of 1938: they used the same sorts/metal types but not not "lift" kasra, resulting in a less clear lines.)) These were made into printing plates in Giza ‒ where they already had ex­perience with printing maps in off­set. Printing was also done there.
Type printing is a letter­press process. The types/sorts leave small inden­ta­tions on the paper: the types/sorts press the printing ink into the paper. Offset is plano­graphic: the paper absorbs the ink; you can't find inden­ta­tions. With his eyes, Mohr saw that it was not hand­written. But he does not know that type print can only be recog­nised with the sense of touch (not by vision). And neither did Prof. Dr. Murks. "That's nonsense, instead of elabora­tely type­setting and printing that ONCE, why not have a calli­grapher write it?" This fails to appre­ci­ate the technoid sense of accuracy of the editors of 1924. To this day, there is no one except ʿUṯmān Ṭaha (UT) who is as accurate as the type­setting or the computer.
Two examples to illu­strate.

While UT clearly reads yanhā, the beauti­ful Ottoman hand­writing reads naihā; while the three vowel signs (fatḥa, sukūn, Lang-ā) are clearly in the right order (there is no other way, they are all on top), nūn (perhaps) comes before yāʾ (does the nūn dot come before the yāʾ dots). Inciden­tally, the two "tooth" letters have a tooth or spine in UT, but none in court Ottoman! While there is clearly nothing bet­ween heh (I use the Uni­code name to clearly dis­tin­guish it from ḥāʾ) and alif maq­ṣūra in UT, there could well be a tooth in Otto­man: You only needed to put two dots over it and it would be hetā or something like that.

Second example: wa-ma­lā­ʾi­ka­tihī Whereas in the 1924/5 Qur'an (below) and UT (in the middle) there is a substitute alif-with-madda hovering BE­FORE the tooth above the base­line, in Muṣḥaf Qaṭar (above) there is a hamza-kasra hovering AFTER the change alif-with-mad­da below the base­line, which changes the yāʾ-tooth into a (leng­then­ing) alif. There is nothing wrong with this (sound and rasm are the same, after all), but it is a dif­ferent ortho­graphy and should not be, according to the con­ception of people who do not to­le­ra­te any ap­pro­xima­tion in the Qur'an.
Now the whole of page 3 in com­parison. Giza print and UT: the Amiriyya is more calligraphic than UT, which can be seen in the examples in the right margin.
All in all, UT follows the default. Baseline and clear from right to left. Only in the spacing bet­ween words is it less modern than the Amiriyya (which is why Dar al-Maʿrifa increased the spacing).

Also from page 3 Com­parison of Muṣḥaf Qaṭar and UT. In the first and last examples, Abū ʿUmar ʿUbaidah Muḥammad Ṣāliḥ al-Banki / عبيدة محمد صالح البنكي does not place the yāʾ-dots EXACTLY under the tooth (in the first case because of the close nūn, in the second case without need). Three cases show tooth letters without a tooth. And a cuddle-mīm, which makes its vowel sign sit wrong (for modern readers): the mīm is to the right of the lām, but the mīm vowel sign is to the left, because the mīm is to be pro­nounced after the lām. So it is rightly "wrong".
Before I stop (for today): a map of Cairo 1920, on which I have marked the Amiriyya and the Land Registry with arrows in the Nil, as well as Midan Tahrir and the place where the government printing press is now located. Also the Ministry of Edu­cat­ion and the Nāṣi­rīya, where three of the editors worked.
Everything to the right of the Nile plus the islands is Cairo, every­thing to the left (Imbaba, Doqqi, Giza) not only does not belong to the city of Cairo, but is in another province.

Important: the typesetting workshop and the offset workshop were well connected by car, tram and boat. The assembled pages did not have a long way to go.
The two Arabic texts are the 1924 and 1952 printer's notes, both from the copies in the Prussian State Library, which owns five editions. And here is the very last (un­paginated) page of the original print.
"al-Qāhira" has to wait till the Fifties to appear.