Learning Taiwanese / Hokkien / Minnan / Min

[Dragonbones]

No problem! And welcome to the forum, by the way!

[yuli]

The following quotations are from a post made 7 years ago. Given that they are that old, they do obviously not represent the poster's understanding today, and i bring this up only because it offers a great opportunity to illustrate how one can develop characters for Taiyu.

[...]many pages give the meaning in Chinese characters, but clearly these are not necessarily the characters corresponding to those particular spoken Taiwanese words. For instance, p. 401 gives "tak8 各, 每 each/every." (I converted the diacritical to a numeral.) So which is it, 各 or 每? Neither, I suspect. I'm guessing tak8 is 諸, based on 讀 being tak8

I'd say the obvious choice is "各". Why? In Mandarin, we have "各 ge4 each/every" and "每 mei3 each/every" in contrast to "諸 shu1 all/various" and "讀 du2 read/study/pronounce", and in Japanese we find "各 kaku each/every" and "毎 mai; bai every" in contrast to "諸 sho many/several", and "讀 doku; toku read; study; pronounce" (though nowadays "讀" is written as "読"). Considering the Japanese sound "kaku", "各 tak8 each/every" would make perfect sense, while "諸 tak8 each/every" would be quite a stretch. After all, there is nothing strange about thinking of "kaku", "ge4", and "tak8" being related by way of an "ancestor sound".

"What's the character for chiao2, bird? 隹? 梟?

Well, "chiao2" could be etymologically related to "隹 zhui1" (Japanese "sui"), but it seems more obvious to relate it to Mandarin "鳥 diao3 bird" (the original pronunciation of today's "niao3") and Japanese "鳥 chou bird", and in terms of communication, too, "鳥 chiao2 bird" would be a better choice, as someone else had pointed out.

What I want to say is that in order to decide which kanji to use for a given expression we need to consider etymology and historic sound changes as well as common usage in related languages as opposed to relying on how things are pronounced today.

This is just for the record (since newcomers apparently read, and learn from, old threads).

[Tempo Gain]

[...]many pages give the meaning in Chinese characters, but clearly these are not necessarily the characters corresponding to those particular spoken Taiwanese words. For instance, p. 401 gives "tak8 各, 每 each/every." (I converted the diacritical to a numeral.) So which is it, 各 or 每? Neither, I suspect. I'm guessing tak8 is 諸, based on 讀 being tak8

I'd say the obvious choice is "各". Why? In Mandarin, we have "各 ge4 each/every" and "每 mei3 each/every" in contrast to "諸 shu1 all/various" and "讀 du2 read/study/pronounce", and in Japanese we find "各 kaku each/every" and "毎 mai; bai every" in contrast to "諸 sho many/several", and "讀 doku; toku read; study; pronounce" (though nowadays "讀" is written as "読"). Considering the Japanese sound "kaku", "各 tak8 each/every" would make perfect sense, while "諸 tak8 each/every" would be quite a stretch. After all, there is nothing strange about thinking of "kaku", "ge4", and "tak8" being related by way of an "ancestor sound".

The official character list shows 逐, seems logical. going from the "g" initial of 各 ge4 to the "t" of tak8 seems unlikely, I can't think of another such example.

[archylgp]

What I want to say is that in order to decide which kanji to use for a given expression we need to consider etymology and historic sound changes as well as common usage in related languages as opposed to relying on how things are pronounced today.

Cognates will not necessarily mean the same thing in different languages; for example, 走 [tzou214] means 'walk' in Mandarin and 'run' in Southern Min ([tsau53]). 行 [kia13] ([a] is nasalized) means walk in Southern Min.

The use of cognates in different languages can vary; consider '起':

Southern Min--Mandarin
起來-------------起來
起戌-------------蓋
起性地----------發脾氣
起火-------------生火

So common usage isn't always reliable in determining cognates.

Some words in Southern Min don't have a cognate in Mandarin and vice versa; for example, [sap55 bun13] 'soap' is a loan word in Southern Min that is not in Mandarin. 茶篐 is a term for a kind of soap that is made from tea. This type of soap was replaced by the modern-day variety, which is called 肥皂 [fei35 zau51] in Mandarin. In Southern Min, though, the older term [te33 kɔ44] 茶篐 is used to refer to modern-day soap. Thus, Mandarin 肥皂 [fei35 zau51] doesn't have a cognate in Souther Min.

Min languages don't link up well with the 切韻 phonological system (Middle Chinese). Thus, all but the most old-fashioned scholars believe that these languages evolved directly from Old Chinese and not out of MC. So when trying to explain away differences between Southern Min and Mandarin with historical sound change, you need to start with OC; for example, OC did not have labiodental fricatives (輕唇音). MC labiodental fricatives -- 非，敷，微，奉 -- were bilabial fricatives in OC (重唇音). This explains the regular sound correspondence between Mandarin [f] and Southern Min [p]; such as:
飛 Mandarin [fei], Southern Min [pue]
分 Mandarin [fən], Southern Min [pan]
芳 Mandarin [faŋ], Southern Min [paŋ]

Irregular phonological process that occurred in Southern Min but not in Mandarin (and vice versa) also obscure the relationship between cognates. This usually occurs in set phrases that assimilate a bit. I don't know any complete examples off hand, but I know 不三不四 is such a case.

[yuli]

he official character list shows 逐, seems logical.

Logical in what sense? (This is a question, not a challenge: I don't know Mandarin yet and therefore can't see the logic.)

In Mandarin this is "逐 zhu2 to pursue/to chase/individually/one by one" and in Japanese "逐 chiku pursue/drive away/chase/accomplish/attain/commit".

going from the "g" initial of 各 ge4 to the "t" of tak8 seems unlikely

Not only unlikely, most certainly not the case. My point is that if sound relationships are important one would want to look for a precursor sound- a sound that might have been used 1200 to 1500 years ago and that gave rise to "tak8", "ge4" and "kaku" in different branches of development.

But my main point was/is that, when it comes to deciding which kanji to use for an expression in Taiyu (which is a contemporary creative process), it would seem most obvious to choose kanji that are used in the same position/for the same purpose in related languages (including Japanese) rather than kanji that have the same sound.

To illustrate this with a hypothetical example: assume that in Taiyu the sound for "one" (as a number) was "ong2", and so someone were to suggest to use the character "恩" for it. For a newcomer to Chinese characters that might be OK, but for anybody who can read Mandarin or Japanese this would seem extraordinarily strange - Mandarin: "恩 en1 favor/grace/kindness", Japanese "恩 on grace/kindness/goodness/favor/mercy/blessing/benefit". In that case a much more obvious choice would be the character "一" instead, based on the obvious fact that all related languages use it as one possible character for that meaning - Mandarin: "一 yi1 one/1/single/a (article)/as soon as/entire/whole/etc.)"; Japanese: "一 ichi one".

[yuli]

Cognates will not necessarily mean the same thing in different languages

Of course not.

Some words in Southern Min don't have a cognate in Mandarin and vice versa

Right. And that is a very nice illustration...

Min languages don't link up well with the 切韻 phonological system (Middle Chinese). Thus, all but the most old-fashioned scholars believe that these languages evolved directly from Old Chinese and not out of MC. So when trying to explain away differences between Southern Min and Mandarin with historical sound change, you need to start with OC

A very good explanation...! Similarly, if you compare the sounds of the Ryukyuan languages and Japanese dialects you will need to go back about 1500 years to understand the connection.

What you described here dovetails nicely with what i wrote before ("in order to decide which kanji to use for a given expression we need to consider etymology and historic sound changes as well as common usage in related languages as opposed to relying on how things are pronounced today"): to create a logical character system for Taiyu a variety of linguistic resources needs to be used to make the character set as accessible to speakers of related languages as possible. If, on the other hand, a new writing system is primarily based on contemporary sounds we could end up with something like "諸 tak8 each/every" instead of "各 tak8 each/every" or (the hypothetical) "恩 ong2 one" instead of "一 ong2 one", which would make the communication between users of neighboring languages much more difficult than it already is.

[Taffy]

My underdeveloped spidey-sense for commonly-derived sounds in Mandarin/Taiwanese/Onyomi tells me that ta̍k is not related to gè and kaku. It just doesn't fit. 逐 works slightly better, and is widely used for ta̍k.

On your sound/meaning rationale for choosing characters, volumes have been written on this question - essentially trying to crowbar Han-ji into languages which don't suit them. Taiwanese is not as unsuitable as say Japanese or Vietnamese, but many Taiwanese writers choose to use romanisation instead of (semi-)arbitrarily assigning characters to tricky syllables.

A further problem is that this project to rationally choose characters for Taiwanese has been underway since at least the time of Lian Heng, and getting people to agree on (a) an etymological or "bun-ji" (本字) approach and then (b) a "correct" answer has proven impossible. I would venture to say that the majority of writers in the Taiwanese literature/linguistics field view such a project with deep suspicion.

[yuli]

many Taiwanese writers choose to use romanisation instead

If the government were to decide to go that route, do you think the PRC would feel compelled to use military force against Taiwan?

a "correct" answer has proven impossible.

Very much so. Aside from languages not fitting into neat schemes at the best of times, there are many (open and hidden) biases at play that work against consensus. My own foray into this topic only goes as far as the suggestions i made, based on the notion that there are "less likely" candidates and "more likely candidates" in many cases (and my bias, for the record, is that I am Japanese).

[Tempo Gain]

Logical in what sense? (This is a question, not a challenge: I don't know Mandarin yet and therefore can't see the logic.)

In Mandarin this is "逐 zhu2 to pursue/to chase/individually/one by one" and in Japanese "逐 chiku pursue/drive away/chase/accomplish/attain/commit".

The "individual/one by one" sense is not to far off from "each," there's even a Mandarin term "逐个 [zhúgè] one by one, one after another ..."

There are also quite a few "zh"-initial Mandarin characters with "t" readings in Taiwanese. I don't know how 4th and 8th tone words correspond to Mandarin tones, is there a correspondence?

going from the "g" initial of 各 ge4 to the "t" of tak8 seems unlikely

Not only unlikely, most certainly not the case. My point is that if sound relationships are important one would want to look for a precursor sound- a sound that might have been used 1200 to 1500 years ago and that gave rise to "tak8", "ge4" and "kaku" in different branches of development.

But my main point was/is that, when it comes to deciding which kanji to use for an expression in Taiyu (which is a contemporary creative process), it would seem most obvious to choose kanji that are used in the same position/for the same purpose in related languages (including Japanese) rather than kanji that have the same sound.

To illustrate this with a hypothetical example: assume that in Taiyu the sound for "one" (as a number) was "ong2", and so someone were to suggest to use the character "恩" for it. For a newcomer to Chinese characters that might be OK, but for anybody who can read Mandarin or Japanese this would seem extraordinarily strange - Mandarin: "恩 en1 favor/grace/kindness", Japanese "恩 on grace/kindness/goodness/favor/mercy/blessing/benefit". In that case a much more obvious choice would be the character "一" instead, based on the obvious fact that all related languages use it as one possible character for that meaning - Mandarin: "一 yi1 one/1/single/a (article)/as soon as/entire/whole/etc.)"; Japanese: "一 ichi one".

I disagree with you here.

First, looking back so far is a guessing game. To take the case in hand, "every" is such an abstract term that there's no guarantee of a common ancestry. I think you have to give the phonetic and tonal relationships equal importance to the semantic when choosing a character. "one" isn't a great example as no one would even think of using another character there. If you apply your principle too broadly though, you're throwing a wrench into the phonetic system of Chinese characters as a whole. The official government lists seem to me to do a good job on this front. It's better to use phonetics when the correct character can't be determined.

[archylgp]

The official character list shows 逐, seems logical.

From a historical phonological perspective, the relationship between Mandarin zhú and Southern Min tak8 is regular. 逐 has a 知-series initial, which were all dental stops in Old Chinese. Southern Min still has these Old Chinese stops (although devoiced). Mandarin, however, is far removed from them; 知-series initials became post-alveolar stops before high, front glides; then, they became several different kids of continuant obstruents, eventually coming to what we have today. I think the posters reading this know what happened to the -k.