Why do all romanization systems suck so much?

[chalimac]

So you write "mayor", but "Mallorca", although the origin of the word is the same.

Mallorca is not a name of spanish origin, it's catalan. Hence the use of "ll".

If we can just figure out where these laymen are from . . .

What major language pronounces "c" as [tsʰ]? You can please all, but neither dismiss all.

I can see the virtue in your arguments. As I said romanization is a hard equilibrium. All are loopsided to one side. Romanji, as far as I know, strikes a fine balance for japanese.

The major problem I have with pinyin is not the initials, it's missing vowels.

[roddy]

What major language pronounces "c" as [tsʰ]? You can please all, but neither dismiss all.

Chinese

(I did try to think of something useful to say, honest)

[renzhe]

Most Slavic languages. That's pretty major.

[imron]

Except it's not missing vowels, it doesn't have them to begin with. Finals should be seen as a single unit.

[Guest realmayo]

I heard that pinyin was developed when China was very friendly with the Russians, which explains some of the alphabet decisions.

[renzhe]

Except it's not missing vowels, it doesn't have them to begin with. Finals should be seen as a single unit.

To add to this, even if you added "e" back into "uei", it would still be wrong, because this "e" is different than the "e" in "re" and different still from the "e" in "ben".

All phonetic renderings are approximations.

[chalimac]

Most Slavic languages. That's pretty major.

I've found that Czech and Polish do. You can't really count Russian since it uses Cyrillic.

[renzhe]

You can add southern Slavic languages to that too, Slovenian, Croatian, Bosnian, Serbian and Montenegrin, which all use Latin alphabet (albeit not exclusively). I'm prety sure Slovak does as well.

As for Cyrillic, it can be mapped almost 1-1 to Latin alphabet, at least in the southern Slavic languages (like Serbian), and when you do that, the sound "ц" maps to "c". Bulgarian, on the other hand, romanises "ц" as "ts", but it's "c" in Bosnia, Montenegro and Serbia. Macedonian, which is closely related to Bulgarian, uses "c" when using Latin characters.

In short, Slavic languages using Cyrillic usually transcribe "ц" as "c".

Russian does indeed use "ts".

[chalimac]

To add to this, even if you added "e" back into "uei", it would still be wrong, because this "e" is different than the "e" in "re" and different still from the "e" in "ben".

All phonetic renderings are approximations.

I Agree. But neither is the i in "-iu" as the i in "shì".

For me it's clear than in shuǐ there is a dropped vowel. If we begin leaving sounds unrepresented -however approximatively - then we could just say that the symbol 水 stands for the sound [ʂu̯ei̯]. Sorry for that reductio ad absurdum.

[renzhe]

It's not that absurd.

What you're talking about is a syllabary, like DeFrancis talks about in his book, which is a perfectly reasonable way to write languages phonetically. Some languages are indeed written like that.

Chinese requires less than a thousand characters to be written completely phonetically like that.

The next thing you can do is to split the syllables into initials and finals and write text as a combination of the two. This reduces the number of needed symbols significantly. Then you get bopomofo, which is indeed a phonetic script used for Chinese. Well, bopomofo doesn't always encode finals with a single symbol, either.

The next step is to use the Latin alphabet to represent these. But it's important to understand that you cannot separate -iu (pinyin) into -i-o-u, it's a single sound. It can be written like -iu, or like -iou, or like a drawing of a fish but it's always a unit that cannot be subdivided further.

[Guest realmayo]

we could just say that the symbol 水 stands for the sound [ʂu̯ei̯]

That would be sinosization, not romanization!

[chalimac]

renzhe,

You admitted some posts ago that this same case bothered you and that there was indeed a shortening:

The things that bother me about pinyin:

- uei, uen and iou get shortened to ui, un and iu

It is exactly the same thing that bothers me. All choices in romanizations are simplification and them some are oversimplifications.

Pinyin is here to stay, even more with with Taiwan coming now on board, so maybe I'm beating a dead horse with this post. Anyway, the debate has been lively and documented.

From now on I will use for my notes a romanization called Bacon's cipher in which Wǒ shì Zhōngguórén becomes:

BABBAABBBABAABAAABBBABAAABBAABAABBBABBBAABBABAABBAAABBABABAAABBBABAAABAABAAABBAB

:D:D:mrgreen:

[Lugubert]

If you created a system based on English spelling, then many other groups (think Russians and Vietnamese) would be struggling far more than now.

The optimistic book Catalan in three months uses "imitated pronunciations" based on English sounds. Although I'm fairly familiar with English phonetics, it's still so counterintuitive to my Swedish eyes as to be useless.

[roddy]

Pinyin is here to stay, even more with with Taiwan coming now on board, so maybe I'm beating a dead horse with this post.

I wouldn't say it's a dead horse, but it's a theoretical one. The chances of any changes being made to pinyin are slim, much less to meet any (possible ) benefits to foreign learners. But we'll probably never create a perfect society either, no reason not to try and figure out what one would look like . . .

[querido]

Addressing *only* the goal of encoding pronunciation for foreign learners of Chinese, if I were teaching and had total control over the materials, I would favor a system of non-Roman symbols, such as bopomofo: point to the symbol, play the audio, memorize. If you don't like these symbols, any others would do... except Roman ones: the symbol should mean nothing to you; it is only a pointer to a real Chinese sound. Further, I don't think it would do any harm to learn bopomofo as though it were a set of real Chinese characters, but I think that what some learners do, treating pinyin as though it were a language, is a serious wrong turn. (In other words, since you're going to have to learn a lot of symbols anyway, why not learn real Chinese ones?)

The complaints in these threads are valid: as a system of phoneticizing Chinese whose sounds you have already ingrained, I imagine hanyu pinyin is well designed, but applying it in the other direction clearly causes unnecessary confusion. The solution in practice, given the ubiquity of pinyin, is to be so intimate with the audio of your chosen learning materials, associating the sounds with hanzi themselves, that you automatically eliminate the ambiguities when you see pinyin. Then, I imagine you are using pinyin as the Chinese do, as *pointers to sounds already known*. Massive listening, even before you know what is being said, would be the first step in my opinion.

So I agree with the original poster: Hanyu pinyin is an important convenience because it is already standardized and widespread, but as a means of *learning pronunciation*, any Romanization system is a bad idea. "New Concept Chinese for Children (BLCUP)" (I apologize for mentioning it so often) does not show any pinyin or English for the whole first year- just look at the hanzi and listen to the audio. I agree with this.

[Hofmann]

Ok...no romanization.

чжэ4 ян4 се3 хао3 бу4 хао3?

[atitarev]

Ok...no romanization.

чжэ4 ян4 се3 хао3 бу4 хао3?

zhe4 yang4 xie3 hao3 bu4 hao3

во3 цзюэ2дэ хэнь3 бу2 цуо4 а (wo3 jue2de hen3 bu2 cuo4 a)