Hokkien Minnan Architecture (& language / culture etc)

[SimL]

Hi Niuc,

If not mistaken, Hokkien variants within E-mng/Amoy and Cuanciu/Quanzhou groups all have "-iəng" instead of "eng".

Thanks. So, it’s primarily Zhangzhou/Chiangchew that has “-eng”? It struck me that Douglas (and Barclay) write “-eng”, whereas I was under the impression that Douglas was mostly based on Amoy pronunciation... Perhaps he says something about this in the introduction – I’ll have to have another look.

I would say that "ə" is more obvious in E-mng "-iəng", compared to Tang-ua*/Tong'an (more like "-ing").

I’ve been meaning to send you an article on Sumatran Hokkien, where all words which correspond to Penang Hokkien “-eng” are indeed transcribed as “-ing”. This is an article which Ah-bin sent me.

Hmmm, actually I think "ə" there is due to "ng", rather than meant as real vowel.

I really like this observation. I can imagine a series of steps where “-ing” would be “i” + “ng” and then, (as happened in Tang-ua*/Tong'an) there would be a very slight “ə” at the beginning of the “-ng”, just to make the pronunciation “easier”, and that slight “ə” might gradually become more and more prominent, until it became an essential part of the word, and hence the diphthong “iə” (as in Amoy etc). I'm not saying this is the actual historical sequence, I'm just saying that it's a plausible way that the “-iəng” might have developed.


Hi Andrew,

If you say the phrase chhau3-chho1 or si3-khO1, I don't know whether you agree that the 3-tone does actually have a dip, maybe 54-44 rather than 55-44?

Very good point. For me, it’s high level: negligible dip, and very high. I guess this is what struck Ah-bin when he first met me - the "khuaN3" in "khuaN3-mia7" (fortune-telling) was perhaps where he first observed this.

I'm getting very confused about the actual contours of Penang Hokkien tones now! Particularly as I'm now realising that Penang Hokkien tone-2 is not as sharply falling as Amoy tone-2 (mentioned when I first raised this subject). As I said then, this insight enabled me to realise why Penang Hokkien "ho2" (= "good") sounds so different from Amoy "ho2".

What your question has made me wonder, Andrew, is whether in fact my Penang Hokkien sandhied form of tone-3 might not in fact be the same contour as Penang Hokkien citation tone-2. I never thought this in the past, because I thought of Penang Hokkien citation tone-2 as a (sharply) falling tone, and sandhied tone-3 as a level one. [More on this at the end of the next comment.]

I suppose there is no reason why the a sandhi/running tone must correspond to another citation/standing tone, other than it is more neat.

Yes, some linguists describing “tone-sandhi” have remarked that it’s a “strange co-incidence” that the sandhied tones correspond to (other) citation tones. Indeed, there is no other reason for this than that it is "more neat". Or, put another way, that “the human mind doesn't need to create additional, unneccesary other categories or distinctions”. This in fact happened to me. For years, I've thought that the Penang Hokkien sandhied tone-3 was tone-1, because they were both high-ish level (or so I thought) - my mind simply didn't bother to create another category, even though I now realise that the two tones are quite different.

One might start to venture the opinion that Penang Hokkien at least is an exception to this common pattern - i.e. that in Penang Hokkien, one of the sandhi-tones isn't similar to any of the citation tones.

However (going back to the previous comment), if (as Andrew asked) Penang Hokkien sandhied tone-3 has a slight dip (say 54), and Penang Hokkien citation tone-2 is indeed "high, slightly falling only", then we might have the case that in Penang Hokkien, the sandhied form of tone-3 is in fact the same contour as the citation form of tone-2! [This then restores Penang Hokkien to having the traditional tone sandhi diagram, and we no longer have to have the modified tone-sandhi rule Tone-3 -> Tone-1 (in non-final position).]

Yet the -ieng is used in Beijing Mandarin but not (afaiaa) in Taiwanese Mandarin, but is used in Amoy/Taiwan Hokkien.

My Mandarin teacher was born and bred in Beijing, and the “-iəng” in her pronunciation was one of the first things I noticed. (Until now, she’s the only one I have noticed speaking like this, but that has largely to do with the fact that continuous Mandarin speech is still completely incomprehensible to me, so I would only notice this when someone – like my teacher – is saying single words for us to repeat after her, and I happen to know the pinyin). I would agree (if that is what you are implicitly saying), that the “ə” in many Hokkien variants has an independent origin from the one in Beijing Mandarin. For one thing, there is no obvious (geographical/socio-linguistic) mechanism of transmission of the Beijing one to the Hokkien one. And for another, in the Hokkien case, it’s “-eng” vs “iəng” whereas in the Mandarin situation, it’s “-ing” vs “-iəng” (though Niuc’s example of Tang-ua*/Tong'an is indeed “-ing” vs “-iəng”).

[SimL]

Niuc:

If your friend can do it, I wish he can help to measure my variant's tones too

He and I always have 1,000 different ideas, and there are all sorts of things we are always going to investigate or follow up, but only 1-2 of them actually ever get done. We can certainly do this for you, if ever we get our act together for this particular one...

[SimL]

Oh, I forgot to add that Ah-bin perceives the following three points about Penang Hokkien:
1. Tone-1, a highish level tone, is lower than Tone-1 in other variants of Hokkien, perhaps 44 or 33 instead of 55.

This is very standard in Teochew dialect, where Tone-1's running and standing tones coincide, and both of them are 33.

Thanks tadpole. The one difference still is that Penang Hokkien running Tone-1 is very different from the standing one. If the standing one is 44 or 33, then the running one is a (very) low non-(sharply)-rising, non-(sharply)-falling one: perhaps 22, 21, or 11, as in "hua1-huiN5" (= "garden"). Personally, I'd go for 44 and 22 respectively, but all will be revealed when my friend and I finally do the tone analysis on the computer (if ever)!

[Andrew]

The more I think of it the less I'm convinced that the running tones are constant: e.g. compare what sounds to me as 四箍 si54khO44 and 四碗 si44ua*53 四粒 si54liap4. Am I alone in finding it difficult to pronounce two falling tones together?

It took me a long time to realise that the Beijing tones are very different from the Hokkien tones: the high level (1st) tone is as high if not possibly higher than the falling (4th) tone, whereas in Hokkien the falling (2nd) tone is higher than the high level (1st) tone. We tend to speak Mandarin with the Hokkien tones in Penang, especially noticeable with the falling-rising (3rd) tone, which is pronounced as low/low falling as in Hokkien (7th/3rd).

[Andrew]

PS I have checked Bodman, who has this to say for Amoy:

In the tones distinguished by different pitch levels, that is the high mid and low tones, the isolation values are all slightly lower than the combination values. An example of the two types of high tones is kaû-khô [Bodman uses sandhi values], nine dollars, where the pitch level of khô is slightly lower than the level of kaû. Similarly, for the two mid tones, in the example sā:-uī, three people, the pitch level of uī is somewhat lower than the level of sā:. An example of two low tones is gŏ-tĕ, five pieces, where the pitch level of tĕ is lower than that of gŏ. In this example, whereas gŏ is simply on a low level pitch, the tone of tĕ, while also low, falls off noticeably.

In the case of the falling tone, the degree of the fall is larger in isolation than in combination ... for example, compare the two falling tones in chaì-kuàn, restaurant ...

In other words, most of the sandhi tones do not in fact correspond exactly to a citation tone. His tone diagram can be rendered, in stereotone analysis:

1. 33-44 2. 55-42 3. 43-21 4ptk. 5-2 4q. 54-21
5. 33-24 6. 55-42 7. 22-33* 8ptk. 2-5 8q. 2-5

*slightly lower than the sandhi 33 in 1 & 5.

According to Glossika, Zhangzhou has a lower 1st tone (44) than Xiamen (44/55)

[niuc]

Thanks. So, it’s primarily Zhangzhou/Chiangchew that has “-eng”? It struck me that Douglas (and Barclay) write “-eng”, whereas I was under the impression that Douglas was mostly based on Amoy pronunciation... Perhaps he says something about this in the introduction – I’ll have to have another look.

hi Sim: I think Douglas just used existing Peh-Oe-Ji 白話字 (POJ) system. I don't think he was the inventor of POJ, or was he? The very spelling of 'oe' is inaccurate for 話 'ue7' in E-mng/Amoy Hokkien. 'Oe' sounds like Teochew.

I’ve been meaning to send you an article on Sumatran Hokkien, where all words which correspond to Penang Hokkien “-eng” are indeed transcribed as “-ing”. This is an article which Ah-bin sent me.

Thanks, I'd be happy to read it. There are at least two groups of Hokkien in Sumatra. The Northern Sumatran Hokkien is very similar to Penang type, while the Eastern Sumatran Hokkien (Riau) is similar to Melaka/Johor type.

From the grammatical role of many YangShang-tone words, I am developing a sense that the original Old Hokkien (older than Middle Hokkien which stabilized in Southern Song dynasty) was a tone-less language.

hi Tadpole: Is this particular for Old Hokkien only or also for Old Chinese as a whole?

...whereas in Hokkien the falling (2nd) tone is higher than the high level (1st) tone.

hi Andrew: This is not the case for my variant. Our 2nd tone has low pitch, while the 1st tone is similar to Mandarin. I notice that 2nd tone in some Hokkien variants is similar to Mandarin's 4th, is this the case for Penang too? For my variant, it is the 8th tone (similar to Mandarin's 4th).

[SimL]

The more I think of it the less I'm convinced that the running tones are constant: e.g. compare what sounds to me as 四箍 si54khO44 and 四碗 si44ua*53 四粒 si54liap4. Am I alone in finding it difficult to pronounce two falling tones together?

Indeed, I think this is the point which tadpole is trying to make about the "correct" use of the term "sandhi-tone". This could/should be used for the "smaller" differences of a "conceptual" tone contour, when it is in different contexts, as the 3 examples above. That's why he prefers the terms "running" and "standing" to capture the differences between phrase/term-non-final and phrase/term-final for what we traditionally call "sandhi" and "citation" tone. These "running" and "standing" tones can then have (in turn) sandhi-forms, depending on their context.

It took me a long time to realise that the Beijing tones are very different from the Hokkien tones... We tend to speak Mandarin with the Hokkien tones in Penang...

You mention another factor which adds a lot of complications to our ability to think about this subject. Namely, we think we are comparing our Hokkien tones to "Mandarin" tones. But these so-called Mandarin tones are the ones which occur in our own heads: the way we've been taught to speak and perceive Mandarin. And this was before the globalization of the 1990's, the rise of PRC film industry, the presence of huge numbers of PRC students in other parts of the world, mass tourism to China, cheap telephony, internet, skype, youtube clips, etc). So, all the models we had when we were forming our conception of Mandarin were based on forms of Mandarin which are/were heavily influenced by Hokkien (or another southern dialect), or by Taiwanese Mandarin, etc. Even when we nowadays hear standard northern PRC Mandarin, we probably still map them to the framework we learnt in our youth. So, I think this really complicates our ability to make accurate statements about how "Hokkien tones" compare to "Mandarin tones".

Anyway, I going to put this subject (for myself) on hold for a while, until I can draw up a complex table of "standing" and "running" tones (in different contexts) and then do the tone analysis with that friend of mine.

I'll share with the Forum when this finally gets done.

[SimL]

hi Sim: I think Douglas just used existing Peh-Oe-Ji 白話字 (POJ) system. I don't think he was the inventor of POJ, or was he?

Hi Niuc,

No, I don't think he was the inventor of it. I'll have a look at what Henning Kloeter says about this in his book "Written Taiwanese". I'm sure he would discuss the history of POJ as part of the introductory material to his main topic.

There are at least two groups of Hokkien in Sumatra. The Northern Sumatran Hokkien is very similar to Penang type, while the Eastern Sumatran Hokkien (Riau) is similar to Melaka/Johor type.

Ah, I see. This article is on Northern Sumatran Hokkien. You're from Bagansiapiapi (http://en.wikipedia.org/wiki/Bagansiapiapi) right? I imagine that your variety is the Eastern Sumatran type then.

[SimL]

Nowadays E-mng pronunciation of "good" 好 'ho2' sounds similar to my pronunciation of "fire" 火 'her2'. That was not the case when Douglas wrote the Amoy Vernacular Dictionary.

Not only Amoy, it would seem. In the second of your instruction links (in another thread): 视频: 闽南语教程1 (video lessons)
http://v.youku.com/v_show/id_XNjUxODYzODA=.html, one can hear the speaker say "ho2" (= "good"), "bo5" (= "don't have, not present"), "koh4" (= "still, additional") as "her2", "ber5", "kerh4". This seems to me to be a Taiwanese site, so it would appear that Taiwanese Hokkien is doing this too. Strange, because sound-change usually in areas with intimate contact, and political / physical barriers usually block the spread of a sound-change.

[niuc]

Hi Sim

Ah, I see. This article is on Northern Sumatran Hokkien. You're from Bagansiapiapi (http://en.wikipedia.org/wiki/Bagansiapiapi) right? I imagine that your variety is the Eastern Sumatran type then.

Yes, Eastern Sumatran type, I was in Bagansiapiapi until my teens.

Not only Amoy, it would seem. In the second of your instruction links (in another thread): 视频: 闽南语教程1 (video lessons)
http://v.youku.com/v_show/id_XNjUxODYzODA=.html, one can hear the speaker say "ho2" (= "good"), "bo5" (= "don't have, not present"), "koh4" (= "still, additional") as "her2", "bo5", "kerh4". This seems to me to be a Taiwanese site, so it would appear that Taiwanese Hokkien is doing this too. Strange, because sound-change usually in areas with intimate contact, and political / physical barriers usually block the spread of a sound-change.

Yes, indeed 無 "bo5" there is closer to "bo5" than to "ber5". It also seems that the male speaker's "-er" is closer to "-o" compared to the female. I have heard those (E-mng/Amoy TV news, if not mistaken) who pronounced "-o" as purer "-er" than in that link. So I think they are not identical.

You are right about political & physical barriers. I suspect that in E-mng (Xiamen) there were at least two variants, one closer to "-o" and the other closer to "-er". Another similar sample is the pronunciation of 豬 as "tu1" (like in Penang), which is neither Ciangciu ("ti1") nor Cuanciu ("ty1"), but of certain village in Tang-ua*/Tong-an (now under Xiamen). So there were at least 2 variants (common "ty1" and not so common "tu1") around Tang-ua* area.

In Taiwanese tv programs, usually I hear "-o" pronounced as (or closer to) "-o", rarely "-er". However, there are many Hokkien variants in Taiwan, which of course are brought over from Minnan area. I think the link (Harvard Taiwanese lessons) is using Taiwanese correspondence of E-mng Hokkien as the standard.