0

I have been messing around with converting a large Devanagari Sanskrit text to Brahmi using a simple mapping function. There is also this table showing how most of the characters map. I found one that doesn't though, the . Right now I'm just mapping it to अ, but please correct me if I'm wrong.

What I'm wondering about though is how to convert Sanskrit Devanagari text to Brahmi. Can I just strip the diacritics and call it a day? Or do I need to be more careful and look at (unicode) character combinations (like letter + diacritic) to map to a single (or perhaps multiple) Brahmi letter(s)? How do I convert Sanskrit Devanagari to Brahmi?

Right now I have this (the mapping is just what I wanted to show in case it's relevant):

var MAP = {
  "ऽ": "\u{11005}",
  "ॠ": "𑀌",
  "\u0905": "\u{11005}",
  "\u0906": "\u{11006}",
  "\u0907": "\u{11007}",
  "\u0908": "\u{11008}",
  "\u0909": "\u{11009}",
  "\u090A": "\u{1100A}",
  "\u090B": "\u{1100B}",
  "\u090C": "\u{1100D}",
  "\u090F": "\u{1100F}",
  "\u0910": "\u{11010}",
  "\u0913": "\u{11011}",
  "\u0914": "\u{11012}",
  "\u0915": "\u{11013}",
  "\u0916": "\u{11014}",
  "\u0917": "\u{11015}",
  "\u0918": "\u{11016}",
  "\u0919": "\u{11017}",
  "\u091A": "\u{11018}",
  "\u091B": "\u{11019}",
  "\u091C": "\u{1101A}",
  "\u091D": "\u{1101B}",
  "\u091E": "\u{1101C}",
  "\u091F": "\u{1101D}",
  "\u0920": "\u{1101E}",
  "\u0921": "\u{1101F}",
  "\u0922": "\u{11020}",
  "\u0923": "\u{11021}",
  "\u0924": "\u{11022}",
  "\u0925": "\u{11023}",
  "\u0926": "\u{11024}",
  "\u0927": "\u{11025}",
  "\u0928": "\u{11026}",
  "\u092A": "\u{11027}",
  "\u092B": "\u{11028}",
  "\u092C": "\u{11029}",
  "\u092D": "\u{1102A}",
  "\u092E": "\u{1102B}",
  "\u092F": "\u{1102C}",
  "\u0930": "\u{1102D}",
  "\u0932": "\u{1102E}",
  "\u0933": "\u{11034}",
  "\u0935": "\u{1102F}",
  "\u0936": "\u{11030}",
  "\u0937": "\u{11031}",
  "\u0938": "\u{11032}",
  "\u0939": "\u{11033}",
  "\u093E": "\u{11038}",
  "\u093F": "\u{1103A}",
  "\u0940": "\u{1103B}",
  "\u0941": "\u{1103C}",
  "\u0942": "\u{1103D}",
  "\u0943": "\u{1103E}",
  "\u0944": "\u{1103F}",
  "\u0962": "\u{11040}",
  "\u0963": "\u{11041}",
  "\u0947": "\u{11042}",
  "\u0948": "\u{11043}",
  "\u094B": "\u{11044}",
  "\u094C": "\u{11045}",
  "\u094D": "\u{11046}",
  "\u0964": "\u{11047}",
  "\u0965": "\u{11048}",
  "\u0966": "\u{11066}",
  "\u0967": "\u{11067}",
  "\u0968": "\u{11068}",
  "\u0969": "\u{11069}",
  "\u096A": "\u{1106A}",
  "\u096B": "\u{1106B}",
  "\u096C": "\u{1106C}",
  "\u096D": "\u{1106D}",
  "\u096E": "\u{1106E}",
  "\u096F": "\u{1106F}"
}

var combining = [
  '\u0900',
  '\u0901',
  '\u0902',
  '\u0903',
  '\u093a',
  '\u093b',
  '\u093c',
  '\u093e',
  '\u093f',
  '\u0940',
  '\u0941',
  '\u0942',
  '\u0943',
  '\u0944',
  '\u0945',
  '\u0946',
  '\u0947',
  '\u0948',
  '\u0949',
  '\u094a',
  '\u094b',
  '\u094c',
  '\u094d',
  '\u094e',
  '\u094f',
  '\u0955',
  '\u0956',
  '\u0957',
  '\uA8E0',
  '\uA8E1',
  '\uA8E2',
  '\uA8E3',
  '\uA8E4',
  '\uA8E5',
  '\uA8E6',
  '\uA8E7',
  '\uA8E8',
  '\uA8E9',
  '\uA8Ea',
  '\uA8Eb',
  '\uA8Ec',
  '\uA8Ed',
  '\uA8Ee',
  '\uA8Ef',
  '\uA8f0',
  '\uA8f1',
  '\u1CD1',
  '\u1CD4',
  '\u1CDb',
  '\u1CDe',
  '\u1CDf',
  '\u1Ce2',
  '\u1Ce3',
  '\u1Ce4',
  '\u1Ce5',
  '\u1Ce6',
  '\u1Ce7',
  '\u1Ce8',
  '\u1CF7',
]

var chars = []
for (var char of text) {
  char = char
    .replace(/\s+/g, '')
    .replace(/[\{\};\:\.,\"\'\(\)।॥\[\]]/g, '')
  if (!char) continue // ignore whitespace
  if (combining.includes(char)) continue // ignore diacritics?
  chars.push(MAP[char]) // brahmi character
}

What could I be doing better?

Note, I am wondering less about the code/technical aspects, and more about the linguistic aspect of orthography to orthography.

7
  • 2
    As a general rule, diacritics are there for a reason. Just because English doesn't use combining diacritics doesn't mean they're not important in other languages!
    – Draconis
    Nov 11 '19 at 4:29
  • Original Quran didn't use them. Nov 11 '19 at 5:16
  • 3
    I can't read Brahmi, but I have never seen the ऽ in the current writing of any of the Indic scripts I can read. Stripping diacritics from Sanskrit Devanagari will render it gibberish. Why don't you do a 1:1 map between the diacritics too?
    – prash
    Nov 11 '19 at 10:16
  • 1
    I should have explained why removing diacritics is a bad idea: these diacritics indicate vowels, and removing them has the effect of replacing all vowels with the default, ǝ.
    – prash
    Nov 11 '19 at 16:48
  • 1
    Unless I'm very mistaken, original Brahmi was also an abugida like Devanagari is, so if by "diacritics" you mean the vowels, then removing them would make it gibberish in either script.
    – LjL
    Nov 11 '19 at 18:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.