Saturday, January 23, 2010

Urdu Indian alphabet order makes best sense

See end - pic is of a proposed Chomsky phonology-derived HIOXian diacritic system.

D: The Hindi alphabet is laid out with a keen sense of articulation.
Letters are grouped by the means by which the sound is made.
Like with like.

The overall sequence of all alphabet has a specific order. The order of the letters depends on the place (in the mouth) where it is pronounced. The sound of the letters in the alphabet start from the back (deep) of the throat and steadily moves upwards & outwards to the front of the mouth (towards lips).
So when one speaks Hindi alphabet, the sound (of the letters) constantly moves outwards starting from the bottom of the mouth and ending towards the front of the mouth. Isn't that fascinating?

And that's not all. Each group of letters above (usually grouped in four), are also arranged in specific sequence. Take first four letters for instance: क ख ग घ.
The same sequence (non-voiced/non-aspirated to voiced/aspirated) follows for:
D: a guide to how to pronounce the
Hindi alphabet.

HIOXian will have a 'letter order' based on the Hindi system.

D: the other pics are of a proposed HIOXian system. It is based on a hybrid of traditional and Chomsky-esque phonology.
I will likely prioritize both English and Mandarin (and so Decimese) in the core letter, so that the diacritic is strictly optional.
I suspect the core figure, the HIOX (H on I on O on X) will contain adequate information for a particular language.
The diacritic will provide additional information. Sometimes this will be redundant with the core figure.
Sometimes the diacritic will contain additional information on timing.

D: Decimese has the core syllable for most vocabulay of CVn - consonant, vowel, nasal consonant.
Initial, mid and final positions.
More complex would be consonant clusters of C plus semivowel.
Dual vowels indicate a diphthong - a "long vowel".
The most complex single-syllable word would be CcVvn.
A slightly more lengthy word would be CcVvcvn.
The top horizontal bars denote time duration. English technically sounds vowels near a mid-word-placed voiced consonant longer. However this is hard to denote with a standard vowel, so 'tis easier to show above said voiced consonant.
So the longest single syllable Decimese word would be CcVvN, as denoted by
1) C - word initial unvoiced but aspirated consonant (PTKFS SH J)
2) c - semivowel consonant. (LRWY)
3) V - vowel (short) (AEIOU)
4) v - diphthong 2nd part (making 3) and 4) a 'long vowel' aeiou as denoted by that or -w or -y
5) N - nasal consonant (M, N, NG)

I may develop 3 systems. Two will be familiar - broad and narrow transcription.
A 'mid-width' version could allow expression of allophones within a language, but without the mind-numbing detail of full-blown IPA.
I may allow 3 vertical levels for the square core figure. High, medium and low.
Our Roman alphabet basically has this.
Mid- aeiou. High - hklb. Low - gjp.
However, HIOXian would not have et al.
I.e. all mid-tier squares would share phoneme-class features.
For example, Chomsky doesn't allow discrete vowel OR consonant categories.
Semivowels occupy the middle no-man's-land.
Concise systems of this nature can capture all English sounds in 15 aspects.
The totality of world languages would require 25.
The basic HIOX figure contains 16 segments. The diacritic contains 7, for 23 so far.
A - - underline (or overline) would push the total to 25.
That's a tight fit!
I think I'll use IPA motif for deep transcription. I.e. a large HIOX figure low and left.
With a smaller HIOX figure high and right.
That would give us ... 16x2 or 32 bar segments to express 25 aspects.

Traditional phonology charts are also unforgiving.
The consonants are laid out in a 7x7 chart, the vowels on 5x5.
A HIOX is 3x3 plus there are diagonal segments.

I would *like* to have the option of all core HIOX figures on the same level.
However, I may need to stick to the original scheme with vowels low and consonants high.
The diacritic would be over and under respectively. I may include the - - opposite the diacritic for additional or even redundant information, to make it visually clear.
Chomsky's theme, though, would refute this neat division between consonant and vowel.

Chomsky's English phoneme features include:
1) vocal
2) consonantal
3) continuous
4) nasal
5) abruptness
6) lateral
7) voiced
8) tensed
9) strident
10) coronal
11) anterior
And front-back -high -low.

The proposed diacritic scheme clearly indicates word boundaries for Decimese syllables and words.
Aspirated-semivowel-vocalic consonantal - vowel- 2nd diphthonged vowel - nasal consonant.
Note. The "T" in "tap" is aspirated - think breathy. T-hhh.
But in Ratan, the T is not aspirated, since it is not word initial.
Ergo, the aspiration diacritic mark in Decimese indicates a word-initial voiceless counterpart consonant to a word-mid voiced consonant.
In English, this reflects rules of sounding out allophones.
Tadan. T- aspirated {T h}. D- voiced, but not aspirated. And so on.
The horizontal diacritic bar segments reflect duration of the sound.
For ease of cursive writing, all word-internal characters may well be drawn with one long arcing line.
I suppose a diagonal or squiggle could be used to denote an implied break.


Special thanks to Adrian for yesterday's contribution!
Never has such a memorable poem been inked on such memorable parchment! <:


Dino Snider said...

Sorry, I meant Hindi not Urdu.
Not sure if there is a difference.

Dino Snider said...

Hmm. I suppose by reducing a standard vowel to only one horizontal diacritic bar, we could introduce gemination. This gemination (syllabic or word-based, like register tonal) could also incorporate suprasegmental prosody considerations. I.e. English stress involves an increased time duration. By reducing a standard vowel, we'd have the option to denote syllable stress on the main vowel also. The diphthonged 2nd vowel (optional) would still be explicitly marked with the centre-verical diacritic bar for vocalic.