From: "Quoc Trung Nguyen" <qt.nguyen@xxxxyyyyzzzzz.de>
To: dchph@yahoo.com
Subject: Vietnamese2020 Suggestions
Date: Tue, 30 Nov 2004 21:42:08 +0100


Dear dchph!

Before I start: I want to apologize that I don't write on vietnamese. I grew up in germany and learnt reading and writing vietnamese by myself. Therefore it is easier for me to write to you in english.

I found your articles about the vietnamese way of writing at vny2k.net. I can't say anything about the roots of vietnamese language but I agree with you that the current vietnamese writing system is not well suited for writing and reading vietnamese. I also agree with you that polysyllabic words should be written as one word and reading in word shapes like in english or chinese is superior to reading letter by letter.

I think a modernized vietnamese script should
- allow reading by word shapes
- be easy to write on paper and on computers
- be able to represent polysyllabic words
- be compatible with the present vietnamese script

Some say that the new style you propose is harder to read and cause misreadings because of ambigous syllable boundaries. I think the problem is that standard quoc ngu already doesn't provide easy to differentiate word shapes. By combining them hard to tell polysyllabic syllables you get a word that is even more difficult to read. I would like to share my ideas about possible ways of solving this problem.

The first I suggest is reducing the number of diacritics, the second is avoiding ambigous syllable boundaries.

Let me show how I reduce the number of diacritics.

The approach bases on a set of 7 vowels. This is not scientific but it helps systemizing the language a lot.

Each base vowel has two representations:

A: a - a(
E: e - ê
I: i - iê
O: o - ô
U: u - uô
X: â - o+
Y: u+ - u+o+

Each vietnamese syllable can be traced down to exactly one of these 14 stems. The rest of the syllable can only be a consonant and/or a half vowel (w sound) before it, and a consonant or half vowel (w or y sound) after it. E.g. Yêu is based on iê, think of it as iê + w. Quê is based on ê, think of it as Gw + ê.

Quoc ngu bases on 11 vowels, that is 6 more than standard latin alphabeth has. My system bases only on 7 vowels, that is only 2 more than standard latin alphabeth has. I assigned new letter combinations to those vowel stems that use diacritics in quoc ngu. The table then looks like this

A: a - ae
E: e - ei
I: i - ie
O: o - ou
U: u - uo
X: ah - oh
Y: uh - uoh

To make use of this system only a few things have to be changed in quoc ngu. The changes are quite easy to remember, after a few minutes you are already able to use it.

a( => ae
ê => ei
ô => ou
â => ah
o+ => oh
u+ => uh
iê, yê, uô => ie, ye, uo

The h in ah/oh/uh is written after the whole vowel cluster. E.g. Mưa is written Muah (and not Muha), Mười is written Muòih (and not Muhòhi). Additionally I use z for d and sometimes dh for đ, because đ and d look too similar.

And how is this better to read?

Remember that you got used to quoc ngu because you have probably used it for years. You will need a certain time to get used to the new system so you can memorize common word shapes. The main advantage of the new system is that the use of diacritics is reduced to tone marks only. Everything that has to do with pronounciation (and I don't mean the tone) is expressed in letters only without any diacritics. The words can be recognized as a whole shape and don't need to be deciphered one letter or diacritic after the other. Compare it with english:

mooh, moe, more is better to read at a glance than mơ, mô, mo.

The same for day, dare vs. đê, đe.
the, though, they are better to read and memorize than thơ, thô, tho.
thoh, thou, tho - don't you also find this better to read than thơ, thô, tho?

Reading in word shapes instead of single letters and diacritics also helps reading polysyllabic words.

How can a polysyllabic word be read comfortably when its syllables causes problems. With a script that provides easy to differentiate word shapes this can be avoided. Even longer words can be recognized at the first glance. Another problem caused by polysyllabic words are the syllable boundaries. For avoiding misreadings caused by ambigous syllable boundaries (cốý, chủnghĩa, sátrạt) I have worked out some rules.

1. before a vowel and before an h, place an apostroph (') to indicate the syllable boundary
2. in the middle of a word ng/ngh in syllable initial position changes to w
3. in the middle of a word tr in syllable initial position changes to j

I personally don't replace tr but if you want 100% misreading-free syllable boundaries you need all three rules. Only then you have a unique representation for each possible syllable boundary. These rules can also be applied on normal quoc ngu without the vowel transcription above. Since quoc ngu has so many diacritics one could easily misread the apostroph as a sa('c tone. In that case I use a dash instead of apostroph, though this is unsatisfactory because dashes require that both sides of the dash are meaningful by themselves. That is not always the case, e.g. nút-radi-ô vs. nút-radi'ô.

I found that when using both, the vowel transcription and the rules for polysyllabic writing, then it is mostly possible to read text without or with only very few tone marks. One could add a tone mark only on the last syllable of a word and the text is still perfectly readable. Since the pronounciation lies completely in the letters the text can be read very well even without diacritics. This is similar to hearing songs where the melody rarely match the inherent tone of the words. This is not that easy quoc ngu, when omitting all diacritics in quoc ngu, not only the tone is lost but also the pronounciation is heavily altered.

Another way to improve readability is noun capitalization. Since I grew up with german language I know about the advantages of noun capitalization. For example if at the first glance you see "ca" then it could be ca or cả, but not cá or cà, because cá or cà would have a capital C. When you know the nouns then you also know that the remaining words are verbs, adjectives, particles, etc. but not things, persons, places.

Here is an example text from your website, once written with quoc ngu and a second time with the changes applied.

Mộtkhi đã nhậnthấy rằng cáchviết hiệnnay không phảnánh đúng một cách khoahọc hiệntrạng của tiếngViệt vì trên thựctế từng từ trong tiếngViệt vẫn còn bị viết rờirạc thành từng tiếng một, bạn hiểu rằng với cáchviết này sẽ làm trìtrệ ócpháttriển của trẻcon khi chúng bắtđầu học đánhvần và nhậndiện chữViệt (xem Ngônngữ và Trítuệ, của Nguyễn Cường) cũng như không tậndụng hết tiềmnăng đadạng của khoa tinhọc hiệnđại.

Mọutkhi đã nhạhntháyh ràeng Cáchviét hiẹnnay khoung phản'ánh đúng mọut Cách-khoahọc hiẹnjạng của Tiéngviẹt vì trein Thụhctéi tùhng tùh trong Tiéngviẹt vãhn còn bị viét ròihrạc thành tùhng Tiéng mọut, Bạn hiẻu ràeng vóih Cáchviét này sẽ làm trìtrẹi Óc-phátjiẻn của Trẻcon khi chúng báetđàu học đánhvàhn và nhạhnyiẹn Chũhviẹt (xem Ngounwũh và Trítuẹi, của Nguyễn Cường) cũng nhuh khoung tạhnyụng héit tièmnaeng đayạng của Khoa tinhọc hiẹnđại.

This script system originally started as a complete alternative to quoc ngu. But I have modified it to be compatible with quoc ngu so both systems can now be mixed within the same text and even within the same word. This is very important for historical and computational aspects.

I hope you can find the ideas useful as a contribution to the development of a modernized writing system.

I would be glad to hear your comments about these suggestions.

Best regards

Nguyen Quoc Trung

-----------------------------