L2/17-377

Proposal for the encoding of three Arabic tanween characters

Eric Muller — Amazon

October 15, 2017

Background

In L2/15-329, Mussa A. A. Abudena proposes 6 characters (1 through 6) for the vowel signs and tanweens, as they are typically rendered in Qalun’s transmission of Nafi’s reading of the Quran.

I agree with the UTC that characters 1 through 3 are just glyphic variants damma, dammatan, and open dammatan. The UTC also suggested that the remaining three characters be represented by the combination of damma or fatha with U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM, and kasra with U+06ED ARABIC SMALL LOW MEEM.

I would like to argue that the representation suggested by the UTC is inadequate, and not just for the typical appearance of Qalun’s transmission of Nafi’s reading, but actually for all calligraphic styles of the Quran. I therefore propose to encode three additional tanween characters, to be used in all transmissions.

Sources

[kfgqpc-h] edition of the Quran made available by the King Fahd Complex For The Printing Of The Holy Quran. It uses Hafs’ transmission of Asim’s reading. This transmission is by far the most popular currently. Unicode text and font available at http://fonts.qurancomplex.gov.sa/?page_id=42 and images downloadable at http://dm.qurancomplex.gov.sa/download/.

[oman] Electronic Mushaf Muscat Calligraphy Project. Also uses Hafs’ transmission of Asim’s reading. Unicode representation is only available punctually. https://www.mushafmuscat.om/.

[kfgqpc-w] edition of the Quran made available by the King Fahd Complex For The Printing Of The Holy Quran. It uses Warsh’s transmission of Nafi’s reading. Available at https://archive.org/details/QuranWarshNarration as images only.

[tunisian] Tunisian edition of the Quran. It uses Warsh’s transmission of Nafi’s reading. Available at https://archive.org/details/QuranWarshAsbahani). If you read Arabic, you can learn about the various tanweeen forms in pages 9-11 of the appendix.

[wics] edition of the Quran, published in 1989 by the World Islamic Call Society - Tripoli - Libya. Uses Qalun’s transmission of Nafi’s reading. Available at https://archive.org/details/Ms7FalGmaHRYaH as images only.

The Quran

Nazalization of the three short vowels fatha, kasra, and damma is ordinarily written using fathatan, kasratan and dammatan (the tanween).

In the Quran, three distinct pronunciations of the tanween are written by different signs:

	fatha	damma	kasra
ordinary vowel sign	[kfgqpc-h] 2:7 [tunisian] 2:7 [kfgqpc-w] 2:6 [wics] 2:6	[kfgqpc-h] 2:6 [tunisian] 2:6 [kfgqpc-w] 2:5 [wics] 2:5	[kfgqpc-h] 2:6 [tunisian] 2:6 [kfgqpc-w] 2:5 [wics] 2:5
ordinary tanween	[kfgqpc-h] 2:182 [tunisian] 2:182 [kfgqpc-w] 2:181 [wics] 2:181	[kfgqpc-h] 2:7 [tunisian] 2:7 [kfgqpc-w] 2:6 [wics] 2:6	[kfgqpc-h] 2:36 [tunisian] 2:36 [kfgqpc-w] 2:35 [wics] 2:35
open tanween	[kfgqpc-h] 2:182 [tunisian] 2:182 [kfgqpc-w] 2:181 [wics] 2:181	[kfgqpc-h] 2:7 [tunisian] 2:7 [kfgqpc-w] 2:6 [wics] 2:6	[kfgqpc-h] 2:36 [tunisian] 2:36 [kfgqpc-w] 2:35 [wics] 2:35
meem tanween	[kfgqpc-h] 2:95 [tunisian] 2:95 [kfgqpc-w] 2:94 [wics] 2:94	[kfgqpc-h] 2:10 [tunisian] 2:10 [kfgqpc-w] 2:9 [wics] 2:9	[kfgqpc-h] 2:99 [tunisian] 2:99 [kfgqpc-w] 2:98 [wics] 2:98

The fragments shown above are the same parts of the text. Qalun’s and Warsh’s transmissions omit the muqatta'at at the beginning of the surah, hence the shift by one in ayah numbers.

Also relevant to the present discussion is the appearance of a small high meem isolated form used above a noon to indicate that it should be pronounced as a meem.

[kfgqpc-h] 2:27
[tunisian] 2:27
[kfgqpc-w] 2:26
[wics] 2:26

Number of occurrences in [kfgqpc-h]

Representation in Unicode

[kfgqpc-h] actually hijacks U+0657, U+065E and U+0656 to represent and display the open tanween, presumably because the text was created before the encoding of U+08F0..08F2.

Representation of meen tanween

Both [kfgqpc-h] and [oman] use <U+064E ARABIC FATHA, U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM> for meem fathatan, and similarly <U+064F ARABIC DAMMA, U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM> for meem dammathan.

[kfgqpc-h] continues the pattern for meem kasratan: <U+0650 ARABIC KASRA, U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM>; whereas [oman] uses <U+0650 ARABIC KASRA, U+06ED ARABIC SMALL LOW MEEM>.

Those representations are not satisfactory.

1. There is an asymetry between the three forms of tanween, since the first two forms are encoded atomically and the meem tanween isn’t.
2. It is quite clear that meem modifies the vowel sign, not the base character, a situation that is not adequately represented by a combining mark on the base character.
3. In the typical shapes used for Qalun’s transmission, the shape of the meem in the tanween is clearly distinct from the shape of meem above a noon. Thus it is akward to use the same character (U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM) in both cases.
4. While the systematic use of U+06E2 ARABIC SMALL HIGH MEEM ISOLATED FORM leads to a uniform pattern, it is problematic to use a character with ccc=230 above, while it is actually displayed below the base. The case of shadda + kasra, where the kasra can be displayed just below the shadda and above the base, is only partially a precedent, as shadda and, more importantly, kasra have fixed combining classes.
5. Conversely, using different characters for the small meem depending on whether is attaches to fatha or damma on the one hand or kasra on the other hand is cumbersome at best.

Proposal to encode atomic meem tanween

Consequently, and given the precedent of the open tanween characters, it seems appropriate to encode three new atomic characters for meem tanween, with properties similar to those of the open tanween characters.

08D0; ARABIC MEEM FATHATAN;Mn;27;NSM;;;;;N;;;;;
08D1; ARABIC MEEM DAMMATAN;Mn;28;NSM;;;;;N;;;;;
08D2; ARABIC MEEM KASRATAN;Mn;29;NSM;;;;;N;;;;;

For the representative glyphs, I suggest to use shapes similar to [kfgqpc-h] (built using the existing font).

Proposal Summary Form