accent with anusvara question

funderburkjim commented 4 years ago

The first part of this question regards the correct digitization (using SLP1) of word ending whose Devanagari representation has both an accent and anusvAra.

The images shown are from the scanned images of Lanman.

Example 1:

Current Digitization: vowel + accent + anusvAra

<s>san . te/ parApA/tamAsata ya/tra yatrA/kAmayanta . a/Ta vA/ iya/M</s>

Example 2

Current Digitization: vowel + anusvAra + accent

01556 <lg><l><s>nAsmE/ vi\dyunna ta/nya\tuH si/zeDa\ na yAM 
miha\maki/radDrA\duniM/ ca .</s></l>

First Question:

The printed representations are similar in both cases.

Which is the correct digitization?

Second Question:

The two examples both have udAtta accent.

Is the digitization rule for udAtta also applicable for the other accent:

e.g., should it be vo/caM\ or vo/ca\M ?

Subquestion: In such Vedic texts, is this 2nd accent correctly referred to as anudAtta ?

scharfp commented 4 years ago

The canonical sequence in SLP1 is to place the accent immediately after the vowel. anusvAra is a phonetic segment represented by a separate character that follows the vowel (with or without accent). Where there is in addition nasalization, which is different from anusvAra and is a quality of the vowel, the nasalization follows the accent. Unicode has not yet solved all the combining sequences for multiple combining characters. Yes, anudAtta is the term for a low pitch generally represented by the horizontal line below in Devanagari and by a backslash in SLP1.

funderburkjim commented 4 years ago

So the right sequence is vowel + accent + anusvara (a/M). Right?

funderburkjim commented 4 years ago

Similar question re candrabindu. Assuming candrabindu is a nasalization, I would expect (from your explanation) that the coding would be is it 'vowel + accent + candrabindu' (a/~) . Right?

If 'vowel + accent + candrabindu' is correct coding, then there is an additional twist regarding the Unicode. By my current transcoding, the visually correct Devanagari form is generated by vowel + candrabindu + accent (see examples in #3).

scharfp commented 4 years ago

Yes

Peter M. Scharf, President The Sanskrit Library scharf@sanskritlibrary.org https://sanskritlibrary.org

On 2 Jun 2020, at 3:47 AM, funderburkjim notifications@github.com wrote:

So the right sequence is vowel + accent + anusvara (a/M). Right?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/funderburkjim/lanman/issues/1#issuecomment-637154708, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6QJFV5SBNSC3DTFYCXASTRUQSG3ANCNFSM4NPKT4JA.

scharfp commented 4 years ago

Yes, that’s right, and Unicode requires the different sequence.

Peter M. Scharf, President The Sanskrit Library scharf@sanskritlibrary.org https://sanskritlibrary.org

On 2 Jun 2020, at 3:54 AM, funderburkjim notifications@github.com wrote:

Similar question re candrabindu. Assuming candrabindu is a nasalization, I would expect (from your explanation) that the coding would be is it 'vowel + accent + candrabindu' (a/~) . Right?

If 'vowel + accent + candrabindu' is correct coding, then there is an additional twist regarding the Unicode. By my current transcoding, the visually correct Devanagari form is generated by vowel + candrabindu + accent (see examples in #3 https://github.com/funderburkjim/lanman/issues/3).

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/funderburkjim/lanman/issues/1#issuecomment-637157292, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6QJFQHM3KGU2OT65I23A3RUQTB3ANCNFSM4NPKT4JA.

scharfp commented 4 years ago

Actually neither sequence shows up correctly when I view V/~ or V~/ in example 3. This is because the display is font-dependent and most fonts have not taken care to implement these sequences using the Unicode code points. Even Siddhanta, Chandas, and Uttara, which are the most comprehensive in implementing conjuncts, use private areas to implement accent combinations, not Unicode code points. Peter Freund’s Prajna font does best with accents and is sued on our accent page, but it is not available free ($25).

Latex does best using the skt package.

Yours, Peter

Peter M. Scharf, President The Sanskrit Library scharf@sanskritlibrary.org https://sanskritlibrary.org

On 2 Jun 2020, at 3:54 AM, funderburkjim notifications@github.com wrote:

Similar question re candrabindu. Assuming candrabindu is a nasalization, I would expect (from your explanation) that the coding would be is it 'vowel + accent + candrabindu' (a/~) . Right?

If 'vowel + accent + candrabindu' is correct coding, then there is an additional twist regarding the Unicode. By my current transcoding, the visually correct Devanagari form is generated by vowel + candrabindu + accent (see examples in #3 https://github.com/funderburkjim/lanman/issues/3).

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/funderburkjim/lanman/issues/1#issuecomment-637157292, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6QJFQHM3KGU2OT65I23A3RUQTB3ANCNFSM4NPKT4JA.

funderburkjim commented 4 years ago

Thanks, Peter!

funderburkjim / lanman