Thursday, March 20, 2014

Gujarati Orthography for the Transliteration of Arabic

Khoja communities such as the Ithnashari ("Twelver Shia") and Agakhani use a convention for transliterating Arabic into the Gujarati script. It consists of adding diacritics to Gujarati letters which most closely approximate the Arabic sound being represented. Some of these diacritics are shown below, with color coding added, in an excerpt of a printed version of the Quran in the Gujarati script:

I've written a proposal to encode these characters as combining signs in the Gujarati block of the Unicode standard. I am interested in communicating with users about these and other signs used in Gujarati for similar purposes. Thank you to Iqbal Akhtar for informing me about these signs.

Sunday, January 5, 2014

A "New" Khojki Inscription

I suppose it's sort of a reverse Rorschach Test: whenever I see a blank slate, my mind will invariably conceive of some sort of orthographic imprint to fill it. I admire inscriptions and often wonder about the talented artists who chisel exquisite texts in complex scripts into marble and about those creative souls who render delicate engravings on copper plates and brass urns. I finally got a chance to experience it for myself last month.

At the beginning of December 2013, I was invited to attend an open house at Equinox Studios in the Georgetown neighborhood of Seattle, where the artists in residence had flung open the doors of their studios and invited inquisitive visitors to experience the creative spaces where inspiration meets the workbench. In addition to the industrial performances of an artist swinging a wrecking ball into massive, ornate glass sculptures and the iron-wrights who lit up the night with pyrotechnic displays emanating from giant, intricate lattices and delicate cut-metal trees, there was in one corner of a garage a more silent performance that caught my ears.

As I made my way to that garage, I spotted sandstone blocks stacked into short towers on the ground and people hunched over tables. I curiously peered over the shoulder of a couple working away at one of these blocks, diligently carving the logo of the Seattle Seahawks into the white sand. Beyond the tables was a cordoned-off area where metal workers were feeding wood into a raging furnace. I watched as a pair of workers dressed in protective suits reached into its blazing mouth with a heavy metal yoke and extracted a crucible of molten metal. Another worker collected finished blocks from the hands of eager carvers and placed them into rows on a metal rack resting upon a bed of black sand. At her signal, the bearers of the crucible inched over to the rack and steadily poured the molten steel into each block. I moved along the perimeter and watched another worker crack the blocks that had cooled in order to release the newly embossed metal tiles.

I was engrossed by this process of sandstone etchings being formed into pictographs on steel and nearly forgot about the rest of the open house. Some of my companions wandered onto the other installations as I grabbed a sandstone block, wriggled my way into the empty space at a table next to the couple, who were putting the final striations on their Seahawks design. I pondered what significant text I should want wrought in steel. Luckily, my friend curtailed that process when she exclaimed "Write my last name in Hindi!". I gladly obliged. Calligraphy I've done, but etched text? Never.

For the next 45 minutes I stood there, hunched over my block, sketching her name in pencil, then lightly tracing it with a carving tool, then going over the grooves repeatedly until the depths reached a quarter of an inch. Then the devil of details arrived and he sat on my shoulder as I tried to coax the stone to yield curves that resemble those of inked letter-forms. When I finished, I brushed off my shoulder and handed the finished block to my friend, who ran her finger over the grooved of the etched imprint of her name in reverse. Thumbs up. She passed it onto the metal worker, who sprayed it with a graphite coat and set it on the rack. We watched as the two crucible bearers arrived with a fresh trove of molten steel and poured it onto the carving. I would have to wait until tomorrow to see how it all turned out... it turned out nicely.

The experience at Equinox loomed in my mind. What would I have etched into that block? My friend saved me from writer's, um, block with her intervening request. But, what text was worthy enough for me to see wrought in steel? A few days later it struck me as I was researching variant letter-forms used in Khojki manuscripts while listening to Raageshwari Loomba's rendition of the Ismaili ginan "Aaye Rahim Raheman" by Imam Begam:

My friend, for whom I had made the tile with her name, contacted Alair Wells, a talented sculptress and metal artist, whose studio Tinder Heart Metals is housed at Equinox. I described my idea to Alair and she was excited to help. First, we would have to make the sandstone block. We produced a wooden cast for the block by using four 2" x 4"s and affixing it to a plywood board. Then Alair cut an 8" x 10" block from a pine board and we fixed that to the plywood. We poured Washington grade 7 white silica into the cast to measure the quality of sand we would need. Then after we placed the weighed silica into a bucket, Alair and my friend mixed a catalyst and bonding agent into it, mixed vigorously for minutes, and then poured the preparation into the block. I went to pick up pizza from Stellar. We gorged on a Georgetowner while we waited for the 25 lb. sandstone block to harden.

Shown above is the block and the tools with which I was going to perform the etching. For this project, I chose three implements consisting of a " nib, a " nib, and a needle point:

The first step was to sketch out the Khojki text using a pencil. The text must be etched as a mirror image so that the embossing on the tile will have the correct orientation. I had thought of making a stencil by printing out the Khojki text in reverse and cutting out the letters using a fine blade, but the glyphs of the only Khojki font I have do not possess the sort of modulated strokes I desired for the etching. So, I decided to do a free hand rendering of the text in reverse:

I then performed an initial etching. I held my breath quite a bit during this part. Unlike writing on paper or composing in typesetting software, you really cannot go back and erase an erroneous scratch! It might have been alright if I had extra silica and bonding agent on hand. I lightly scraped the " nibbed implement over the sketch I made:


The first phase of the etching of the upper portion of the block is shown below. The grooves are shallow. I would eventually deepen the routes:


Then time to sketch out the text for the bottom portion of the etching. I could have sketched out the entire text first, but I had a feeling that my palm would just smudge the graphite.

Below is the first run of the etching:

I used the broad nib for the text of the upper portion. It's width produced very pleasant stroke width and modulation that resembled the style of Khojki I wanted to emulate.

Etching the tail of a vowel sign:

For the bottom portion, I used the medium nib.

I used the needle point to take care of the small details, like the dots and the terminals of letters:

Working on the details:

I was concerned about the size of the dots. I feared that if I spaced them too closely, then the molten metal might just obliterate the sandstone in between and I'd end up with a blob instead of three dots:

 Here's the before picture of the silica block:

and the after picture:

And a close up of the text:

On the morning of New Year's Eve, I returned the finished block to Alair, who did another pouring in Tacoma that evening. This is how it turned out:


When I was reading a folio from a printed Khojki book earlier in December, I began to wonder if I could reproduce the letter-forms that Laljibhai Devraj had cut in Germany in 1903 for the first ever Khojki metal types, which he used at his Khoja Sindhi Printing Press in Bombay. At first I thought it might be nice to do an etching of "Aaye Rahim Rahman", the ginan which inspired me to think about a Khojki etching in the first place, but that seemed a bit ambitious. After all, I had only done one etching. So, I thought of the next best thing I would want to etch in Khojki...

As I developed my proposal for encoding Khojki in Unicode in 2009, I was contacted by Irfan Gowani, who was enthusiastic about the future encoding. We spoke intermittently, but I did not meet Irfan until I returned to Seattle in 2013. During the past year Irfan and his wife Shelina have become wonderful friends of mine. They are inspirational, creative, and lovely people. And they are damned good cooks. One evening, a few days before Christmas, as I left their home after dinner Shelina handed me two jars of homemade quince and fig jam, sourced from the fruit of their own trees. Apart from the kababs they feed me, this was yet another testament of their giving nature. It got me to thinking: what in the world could I offer them as an expression of my friendship?

What else could more uniquely express my gratitude for their presence in my life more than an 8" x 10" x ¾" metal plaque weighing 17 lb, which bears the names of their family embossed in Khojki by my own hands? I have yet to present it to them. But, I have a feeling that whenever they'll need to move it, they'll surely remember me!

Merry Christmas and Happy New Year, Irfan and Shelina!
āye rahem rahemān, ab to rahem karoge,
āye rahem rahemān, ab to rahem karoge.

ejī tana mana dhana guru ne arpaṇa kīje
to gināne gināne ginān, ab to rahem karoge.

ejī dāna sakhāvat har dam kīje,
to dāne dāne dān, ab to rahem karoge.

ejī saba ghaṭ ekaja rahemān kīse,
to śāne śāne śān, ab to rahem karoge.

ejī kahet īmām begam merā pīr hasan shāh,
īmāne īmāne īmān, ab to rahem karoge.

Thursday, December 12, 2013

A distinct Khojki 'qa' or variant of 'ka'?

The following excerpt from a printed Khojki book contains two forms of ka. The ka highlighted in blue is the representative glyph of U+11208 KHOJKI LETTER KA (to be published in Unicode 7), while the red ka is the alternate form.

Here, the red ka is used in the word kalām, while the blue ka is used in the word kalā̃m (lit. kalām), both are representations of the Arabic word کلام kalām "composition, work"; the anusvāra in kalā̃m is not semantically significant.

The alternate form is used in other sources for representing ka, suggesting that the red ka is a glyphic variant:

However, it also appears in other source, where it is used for representing qa (the Arabic ق qāf), suggesting that it may indeed be a distinctive letter, a *KHOJKI LETTER QA:

I have not yet seen this work, so I do not know the context in which the letter is used for qāf.

The shape of KHOJKI LETTER KA is fairly uniform across printed materials and manuscripts. It is of interest that both forms occur in the same document, as shown in the first image, and in such close proximity. The excerpt is from a printed book and is likely a faithful reproduction of a manuscript. Analysis of several manuscript sources is necessary in order to determine whether the red ka is a glyphic variant of KA or a distinct Khojki letter QA.

Friday, November 15, 2013

Siddham Headstroke: Abode of the Inherent Vowel?

The below excerpt from the Shittan Hidenki [悉曇祕傳記] (Taishō Shinshū Daizōkyō, vol. 84, no. 270) of Shinpan [信範] shows an analysis of independent vowel letters, their dependent forms, and the combinations of these dependent forms with the Siddham consonant letter KA:

It also shows a horizontal stroke associated with Siddham vowel letter A that correlates with the dependent forms of the other vowel letters. In Japanese, this stroke is known as the ア点 a-ten "a mark" and is considered to be an elongated form of the 命点 myō-ten "life mark". The myō-ten is the initial brush stroke used in the writing of all Siddham letters. The a-ten is produced by extending the myō-ten to the right. It is not a true vowel sign; it is the headstroke of each consonant letter.

My unconfirmed hunch is that the Siddham analysis of the a-ten arose as a way of explaining the inherent vowel /a/ possessed by every consonant letter. All Indic vowels have both independent and dependent forms, except for the letter A, which has only an independent form. The a-ten raises several philosophical questions regarding the sound and forms of letters: How does one capture this inherent sound, which is part of the phonetic identity of each consonant letter, but which is graphically unmarked? Is it contained somehow in the letter form? If so, where in the glyph does it reside? As it represents the initial brush stroke used for writing the vowel letter A, it may be said to contain the graphical and phonetic essence of the letter.

I raise the matter of the Siddham myō-ten and a-ten because it is significant from an ideographic perspective, as are the other elemental strokes identified in pedagogical texts. I am currently investigating the potential of encoding these elemental strokes, which I mentioned in my Siddham proposal as being out of scope for the basic encoding. I welcome any information on the myō-ten from users familiar with its philosophical interpretations and its use in Siddham pedagogy.

Tuesday, November 5, 2013

Siddham nukta: A Curious Innovation

The combining sign ़ NUKTA is used in Indic scripts for transcribing sounds for which distinct characters do not natively exist in a writing system. The name of the character is derived from the Arabic word نقطة nuqṭah (simplified in Indic romanization as nukta) "dot". The NUKTA is used in Devanagari and related scripts of northern India for expressing sounds not commonly used in Indo-Aryan languages, mostly those that originate from Arabic and Persian. It is also used natively in scripts such as Bengali and Tirhuta for distinguishing between characters that have nearly identical graphical structures. Although now used quite frequently in modern Indic scripts, the NUKTA is not part of the traditional character repertoire based upon the representation of Sanskrit phonology. As one might expect, the NUKTA is not attested in historical Siddham materials as the usage of Siddham is largely restricted to the representation of Sanskrit.

I was, therefore, surprised when, during research for my proposal to encode the script in Unicode, I stumbled across the following sample at Mandalar showing usage of NUKTA in Siddham text as :

The above excerpt provides the translation or transliteration of English "tattoo" in Sanskrit, English/Hindi, and Japanese. The Sanskrit  वेध vedha does not truly translate as "tattoo", but carries the connotation of a "piercing", "puncturing", or "perforation" (√ व्यध्); the English/Hindi ततू tatū is simply a transliteration of English "tattoo" (which also might be rendered टैटू ṭaiṭū, using retroflex letters instead of dentals); and इरेज़ुमि irezumi, which is the transliteration into Siddham of the Japanese word for "tattoo", 入れ墨 irezumi.

Of interest is the usage of NUKTA in writing the word इरेज़ुमि. In general, the NUKTA is written with a letter that has the closest phonetic proximity to the target sound. Here, the NUKTA is combined with SIDDHAM LETTER JA (/ʤ/) for representing /z/. When NUKTA occurs with a consonant to which a vowel sign is attached, then it is ordered in encoded text immediately after the consonant and before the combining vowel sign: . The positioning of the NUKTA with regard to the base letter depends upon the shape of the letter and the presence of any below-base vowel signs.

The Mandalar site has a chart that shows the usage of NUKTA with other consonants for representing fricatives and other sounds:

The introduction of the NUKTA in Siddham is a surprising innovation. I have yet to investigate the issue fully, but at first glimpse the usage of NUKTA suggests that modern users of Siddham seek to extend the script by adopting features used in Indic scripts. Such borrowings show that Siddham is quite alive in Japan and that its usage is continually being extended beyond traditional contexts. Indeed, the Mandalar site displays these innovations under the banner of 現代悉曇 gendai shittan or "modern Siddham":

In order to support such usage, I proposed the character for encoding as ़ U+115C0 SIDDHAM SIGN NUKTA. It corresponds to characters such as ़ U+093C DEVANAGARI SIGN NUKTA and possesses the same properties.

I am curious to know the history of NUKTA in Siddham and the rationale of the gendai shittan user(s) who first used NUKTA in the script. Was the idea motivated by the usage of NUKTA in modern Indic orthographies? Was it influenced by Unicode? Does the 'dot' have ideographic interpretations?

Monday, November 4, 2013

Siddham myō-ten: The Essence of a Character?

A short downward sloping horizontal stroke is shown in some historical and modern Siddham handbooks in connection with SIDDHAM LETTER A. In Sacred Calligraphy of the East (1981), John Stevens calls it a 'variation' of the vowel letter A:

In the excerpt, below, from the Zusetsu Bonji [図說梵字] (1974) of Kijun Tokuzan [徳山暉純], the stroke is shown as a form of the vowel letter A, on par with the dependent forms of the vowel letters A, I, and II.

Giryū Kodama [児玉義隆] also shows it as a form of A in his Bonji Hikkei [梵字必携] (1991):

Although it is correlated with Siddham vowel letter A, it is not a true vowel sign, but represents the initial brush stroke used in writing the vowel letter. Below is another excerpt from Tokuzan (1974), which shows the nine brush strokes used for constructing the vowel letter A, with the initial stroke highlighted:

The same stroke is used in the creation of all Siddham letters. The stroke is shown below in the forms for Siddham letters A, KHA, HA, RA, VA:

This stroke is known in Japanese as 命点 myō-ten "life mark". It corresponds to the Chinese basic stroke 點 diǎn "dot'", which is encoded in Unicode as ㇔U+31D4 CJK STROKE D.

I have not been able to conduct much research into the meaning of Siddham myō-ten, but my unconfirmed hunch is that it embodies the phonetic power of a character.

the philosophy behind the myō-ten arose as a way of explaining the absence of a mark for the inherent vowel. All Indic vowels have both independent and dependent forms, except for the letter A, which has only an independent form. This raises several questions from a philosophical angle: How does one capture this inherent sound, which is part of the graphical structure of each consonant letter, but which is unmarked? Is it contained somehow in the letter form? If so, where in the glyph does it reside? As it represents the initial brush stroke used for writing the vowel letter A, it may be said to contain the graphical and phonetic essence of the letter.

I raise the issue of the Siddham myō-ten because it is significant from an ideographic perspective, as are the other elemental strokes identified in pedagogical texts. I am currently investigating the potential of encoding these elemental strokes, which I mentioned in my Siddham proposal as being out of scope for the basic encoding. I welcome any information on the myō-ten from users familiar with its philosophical interpretations and its use in Siddham pedagogy.

Saturday, November 2, 2013

Siddham Ornaments: Beyond Punctuation

Siddham has several characters that are used in manuscripts for marking the end of text sections. Some of these are shown in column below:

Palaeographically, these section marks are similar to characters in other Indic scripts, such as 𑁍 U+1104D BRAHMI PUNCTUATION LOTUS, which are ornamental marks. In general, these marks do not possess phonetic values or semantic meaning beyond their function as terminations. However, as the above excerpt shows, in the Japanese analysis of Siddham, these marks additionally represent the syllable "aṃ".

The extent sources contain several other ornaments beyond the three shown above. My research identified at least 14, which I proposed for inclusion in the Siddham block:

The section marks may be classified into five groups based upon their graphical structure.

In addition to their representation of "aṃ", according to certain Japanese Buddhist traditions, these marks have esoteric connotations that offer insights into the textual passages after which they are written. Moreover, the graphical structure indicates other philosophical meanings, of which I am not fully aware or to which I am not privy! These characters would be considered ornaments or glyphic variants of a basic set of section marks according to the character-glyph model of Unicode, but on account of the semantics they possess beyond their function as punctuation, they have been encoded independently.