Also we’d keep in mind that the human filter is a very powerful non-causal system. Meaning that in additional to characters already read in the past, it “peaks ahead” of the current character. So the human filter is a system whose output (i.e., comprehension) depends on future inputs(i.e., next couple of characters) and the past and current inputs(previous characters). We’d want to recognize this functionality and take advantage of it. Make it useful in our design.
（＊＊this is exactly why たくさん is “tak san” (2-syllalbes) to the native Japanese speakers. The perceived shortcomings of kana are common misconceptions. I think I could still introduce the core principles to you even though you’re working strictly on a Roman Alphabet system)
[quote=“sofun”]Also we’d keep in mind that the human filter is a very powerful non-causal system. Meaning that in additional to characters already read in the past, it “peaks ahead” of the current character. So the human filter is a system whose output (i.e., comprehension) depends on future inputs(i.e., next couple of characters) and the past and current inputs(previous characters). We’d want to recognize this functionality and take advantage of it. Make it useful in our design.
（＊＊this is exactly why たくさん is “tak san” (2-syllalbes) to the native Japanese speakers. The perceived shortcomings of kana are common misconceptions. I think I could still introduce the core principles to you even though you’re working strictly on a Roman Alphabet system)[/quote]
as stated before, this system isn’t just for native speakers, and besides, native speakers of Seediq and other Aboriginal languages already have a system superior to Kana, it’s not likely they are going back unless there are incentives of doing so. I’m hoping that having a system that is consistent across the island will be that incentive.
Ultimately, whether a system is good or bad depends on whether the native speakers can identify with it or not. It’d be a mistake not to design from the perspective of the native speakers.
What i see coming out of sofun’s proposal: kana is not inherently unsuitable for (at least some) non-Chinese native languages - that much i agree.
But if one were to use kana well enough, the character set would need to be expanded and in some cases new usage rules would have to be introduced, that would be different from Japanese. The Arabs have introduced new symbols into their alphabet in recent history, so it can be done, i am sure. I suspect, however, that for practical and political reasons kana will not be widely chosen. Using something based on Latin script will probably win out alone on account of the practical advantages of using it in the given electronic communications infrastructure.
Another point, for hansioux: recent posts by sofun. I think we have talked about this before in another thread: if what you are aiming at turn out like a kind of IPA specifically tailored to encompass the native languages spoken in Taiwan, with uncountable diacritics or special (non-Latin) symbols, it may be too “costly”. So why not use a relatively limited set of characters and use them differently in different regions, comparable to what happens in the UK, where you can see English and Welsh written side by side, in Canada, where you see English and French written side by side, in Switzerland, where you can see German, French and Italian (let me omit the minor players here) written side by side, in Belgium, where you can see French and Dutch (as well as languages related to one of either, such as Walloon or Fleming) written side by side, and so on.
In practice this would mean that one language might use letter combinations like “zh” and “zz” for certain sounds while another might end up using “sh” and “ss” for similar but somewhat different sounds instead. And while many language will, of course, share many letters, the pronunciation will differ from language to language - an example might be “j” that in one languages may be pronounced like German “j”, in another like English “j” and yet another like French “j”. I think i have suggested before that what is important is to get away from phonetic distinctions and to work instead from phonemic distinctions: there is a good reason why Germans no longer distinguish “th” and “t” in writing and why “ö” suffices to represent both [œ] and [ø(ː)]…
We could say that, hansioux would need to think about whether he’s working on a romanization as a learning tool, or as a transliteration standard like pinyin, or whether he’s working on a written language. Or both. So this is very important:
If the native speakers already have a written language using the roman alphabet, how do you keep the newly-designed romanization from creeping into (thereby corrupting) the existing written language? Some decisions must be made.
[quote=“yuli”]What i see coming out of sofun’s proposal: kana is not inherently unsuitable for (at least some) non-Chinese native languages - that much I agree.
But if one were to use kana well enough, the character set would need to be expanded and in some cases new usage rules would have to be introduced, that would be different from Japanese. The Arabs have introduced new symbols into their alphabet in recent history, so it can be done, I am sure. I suspect, however, that for practical and political reasons kana will not be widely chosen. Using something based on Latin script will probably win out alone on account of the practical advantages of using it in the given electronic communications infrastructure.[/quote]
I do believe one can make all that modifications to Kana and make it work reasonably well, but then that would be using Kana just for the sake of using Kana, and not because it is truly suitable for the task. Also, the Japanese have tried that before, and the system they came up with wasn’t entirely accurate when it comes to Aboriginal phonology. In the end, why use and create a new Kana system if it is not going to be compatible the phones the Japanese assigned to Kana?
Also, like previously mentioned, having a Kana based writing system doesn’t negate the need for romanization.
absolutely agree. That’s why the goal is not to have any diacritics or non-Latin symbols. I need the table to see if that is possible.
That would not be very different from the current situation, except it requires the political party blocking the 國家語言發展法草案 to get over their ethnocentric-self and pass the darn bill into law. So that there would be more languages aside from Chinese and Pingyin on street signs.
I’m hoping with this system, there will be Chinese and Unified Romanization of the original location name on street signs, and that romanization will allow everyone to say the place name as intended.
Also, the Aboriginals are pushing to register their Aboriginal names only. While I’m all for it, without a unified system being taught in school, not only would Han Taiwanese not be able to read their names from an Aboriginal’s National ID card, other Aboriginals not of the same Aboriginal ethnicity would also have trouble pronouncing it correctly. I feel these types of excuse will always be used to decline the language rights to the Aboriginals unless a unified romanization exists and is taught in school.
[quote=“sofun”]Always enjoyed yuli’s and hansioux’s posts
We could say that, hansioux would need to think about whether he’s working on a romanization as a learning tool, or as a transliteration standard like pinyin, or whether he’s working on a written language. Or both. So this is very important:
[/quote]
I envision for Mandarin speaking Han Taiwanese it would be a transliteration standard. For Holo and Hakka, it would be part of the writing system. For the Aboriginals, it would be the writing system itself.
[quote=“sofun”]Aboriginal languages have a lot of imported words from Japanese and Mandarin. These words are pronounced “as-is”.
A lot. A huge amount. For example news in Paiwan is しんぶん, and school is がっこう。
How do you incorporate this fact into your design?[/quote]
They will be spelled as how each Aboriginal language pronounces them, I don’t see how that would or should be different from any other loan words in any other language.
For example, several Aboriginal language borrowed the Taigi term 歹人 pháinn-lâng to refer to all Han ethnicities. However, many of them pronounces the word as bailan, and that’s exactly how it should be spelled in Aboriginal languages.
[quote=“hansioux”][quote=“sofun”]Aboriginal languages have a lot of imported words from Japanese and Mandarin. These words are pronounced “as-is”.
A lot. A huge amount. For example news in Paiwan is しんぶん, and school is がっこう。
How do you incorporate this fact into your design?[/quote]
They will be spelled as how each Aboriginal language pronounces them, I don’t see how that would or should be different from any other loan words in any other language.
For example, several Aboriginal language borrowed the Taigi term 歹人 pháinn-lâng to refer to all Han ethnicities. However, many of them pronounces the word as bailan, and that’s exactly how it should be spelled in Aboriginal languages.[/quote]
I see. It makes sense. Once it becomes a part of the vocabulary, the definitions may change. It’s totally up to the native speakers.
Just updated the chart with Rukai, Saisiyat and Tsou phonology.
Rukai adds 1 new consonant and 0 new vowels to the Holo-Hakka-Mandarin-Amis-Atayal-Paiwan-Puyuma phonology chart.
The new consonant is:
/θ/
Saisiyat adds 0 new consonants and 1 new vowel to the Holo-Hakka-Mandarin-Amis-Atayal-Paiwan-Puyuma-Rukai phonology chart.
The new vowel is:
/œ/
Saisiyat also has a long vowel marker
:
I am wondering how that should be addressed. If long vowels are to be annotated with the same vowel twice, then we cannot use double vowels to annotate vowels.
Tsou adds 3 new consonants and 1 new vowel to the Holo-Hakka-Mandarin-Amis-Atayal-Paiwan-Puyuma-Rukai-Saisiyat phonology chart.
The new consonant is:
/ɽ/
/ɓ/
/ɗ/
The new vowel is:
/Ɨ/ Spread Closed Central Unrounded, dialects.
So right now I ave a total of 60 consonants and 18 vowels and 9 tones.
[quote=“hansioux”]
Also, the Aboriginals are pushing to register their Aboriginal names only. While I’m all for it, without a unified system being taught in school, not only would Han Taiwanese not be able to read their names from an Aboriginal’s National ID card, other Aboriginals not of the same Aboriginal ethnicity would also have trouble pronouncing it correctly. I feel these types of excuse will always be used to decline the language rights to the Aboriginals unless a unified romanization exists and is taught in school.[/quote]
But Han Taiwanese will never be able to read the Aboriginal’s name, because the phonology does not exist in Chinese. Just knowing the letters does not make you able to read them.
If that were the case, no Non-Chinese would have problems reading Chinese names. But they have. Because they do not know the letter-phonem association (e.g. q = [tɕʰ]) or just because the sound does not exist in their language (for example for me the r sound). And I strongly doubt that Han Taiwanese would develop some interest in aboriginal languages, would they?
Of course, this makes the whole aboriginal names issue quite problematic. I have a hard time imagining how they can practically put only their aboriginal name (in latin letters) on the ID card or use it in daily life when some 90% (don’t know the exact number) have no idea about aboriginal languages. People whose native language uses the latin alphabet can pronounce it somehow (which is then always wrong), but I guess Taiwanese cannot do that.
A solution would be to annotate the reading in Zhuyin. I think Zhuyin is the closest thing to an alphabet more or less everybody in Taiwan can understand. Of course this would again need some modifications. And Zhuyin would then not be something for the name itself, but only for the reading, making non-aboriginal people more or less be able to pronounce the name.
[quote=“Hellstorm”]
A solution would be to annotate the reading in Zhuyin. I think Zhuyin is the closest thing to an alphabet more or less everybody in Taiwan can understand. Of course this would again need some modifications. And Zhuyin would then not be something for the name itself, but only for the reading, making non-aboriginal people more or less be able to pronounce the name.[/quote]
First, changing into Zhuyin would not make the phonology issue go away. If you look at the chart I’ve compiled so far, you would see a system designed to annotate Mandarin would have to under go a huge amount of modifications to just represent all these new phones. After modifications are made, Mandarin only speakers would still have no idea how these new Zhuyin characters represent. So the concept of Zhuyin somehow magically enables Mandarin speakers the ability to read Aboriginal names better than romanization is probably false.
I would imagine without having the new system taught in school, even Mandarin speaker people would have more success of pronouncing an Aboriginal name with romanization any same weird new Zhuyin symbols they’ve never seen. Try making Mandarin speakers read Holo and Hakka Zhuyin extensions, see how far most people get, and you;ll see what I am talking about.
Zhuyin looks horrible as a written language and takes up a log of space for some reason. So imagine fitting an Aboriginal’s name with 10 syllables (plus special small script characters to label consonants with no vowels) on the National ID card.
This is what Zhuyin looks like annotating just 6 syllables in Taigi:
一枝草一點露。
ㄐㄧㆵ̍ ㄍㄧ ㄘㄠˋ，ㄐㄧㆵ̍ ㄉㄧㆰˋ ㄌㆦ˫
Tsi̍t ki tsháu, tsi̍t tiám lōo
[quote=“hansioux”]
First, changing into Zhuyin would not make the phonology issue go away. [/quote]
No, of course not, but I think it is utopic to believe Han Taiwanese will be at sometime magically able to pronounce all new phonemes. If they cannot speak the language, how would they be able to do that? Heck, I cannot even pronounce English perfectly, even though I more or less speak it fluently! So the goal is rather something which is close to the original pronounciation. Something like the Japanese do it with Katakana. Of course it is horribly wrong, but at least it is somewhat related to the original phonetic. Better than to not being able to pronounce it at all, imho.
[quote=“hansioux”]
Zhuyin looks horrible as a written language and takes up a log of space for some reason. So imagine fitting an Aboriginal’s name with 10 syllables (plus special small script characters to label consonants with no vowels) on the National ID card.
This is what Zhuyin looks like annotating just 6 syllables in Taigi:
一枝草一點露。
ㄐㄧㆵ̍ ㄍㄧ ㄘㄠˋ，ㄐㄧㆵ̍ ㄉㄧㆰˋ ㄌㆦ˫
Tsi̍t ki tsháu, tsi̍t tiám lōo
See how it is already considerably longer?[/quote]
Yes, I know how Zhuyin is longer. But Zhuyin is not designed to be written linear, but vertically stacked in syllables. That makes it much shorter.
Well, it was just an idea, because Taiwanese have a hard time with latin letters. Zhuyin is more or less the only thing they know of a alphabetic writing. I know that Zhuyin is ugly as hell and no well designed script.
Yes, but this is what happens when any non-English name is used on a US driver’s license. The police officer checking idea isn’t going to have any idea how to read “Feng Xiaoyue” any better than a Taiwanese Han person can read “Nolay Piho.”
I propose a dual-track system of using both Mandarin characters and Latin letters to transcribe a name: Nolay Piho / 諾萊·彼河 or something like that.
[quote=“Hokwongwei”]
Yes, but this is what happens when any non-English name is used on a US driver’s license. The police officer checking idea isn’t going to have any idea how to read “Feng Xiaoyue” any better than a Taiwanese Han person can read “Nolay Piho.”[/quote]
The difference is that a Western police officer uses latin letters in his daily life and can pronounce something. I have a hard time imagining a Taiwanese police officer pronouncing it.
[quote=“Hokwongwei”]
I propose a dual-track system of using both Mandarin characters and Latin letters to transcribe a name: Nolay Piho / 諾萊·彼河 or something like that.[/quote]
But isn’t that exactly what the aboriginals do not want?
Well, difficult problem. I do not venture out having a good solution.