All roads lead to -ngr

The email discussion on -ngr started soon after the claim was made, on these pages, that -ngr was a sublime consonant. In a spasm of curiosity, a member of the Beijing Sounds staff had written to phoneticist John Wells to try to get a better understanding of just what that érhuàyīn was doing — both in the nasal-oral cavities and in the syllable. John Wells CCd several other eminent linguistics commentators, and what results is explication of some of the most pressing Beijing-R questions ever to cross the minds of Olympics spectators. The details are elaborated in the section below.

But I’ve been sitting on that conversation since March. The problem is the trepidation that comes with explaining any phenomenon. The ngr could end up a victim of clinical description via the same process that transformed an epic event in which “the sun…seemed terribly darkened over like sackcloth of hair” into, well, a solar eclipse: neat, but understood. Who doesn’t worry that science is the death of poetry, regardless of how convincing Feynman is?

Even worse, it might turn out the ngr is really nothing after all: a microphenomenon explained by some more important general rule, a motley collection of events not worthy of being grouped together.

But the words of Steven Pinker bring back some hope for the ngr. Paraphrasing Kripke, he offers the tantalizing prospect that an idea, once expressed in a word, actually becomes something and will never be made any less “real” no matter what explanation the people in white coats might give.

Kripke argued for yet another possibility, one that most philosophers had never even considered: knowledge that is a posteriori (discovered after the fact), but necessary. The discovery that the Morning Star and the Evening Star were the same thing (Venus) was a posteriori. But once it was discovered, it was a necessary truth — there is no possible world in which the Morning Star and the Evening Star refer to different things…. Kripke’s argument is an attempt to clarify what we are logically committing ourselves to when we use proper names and names for natural kinds. We are, surprisingly, committing ourselves to a certain class of logically necessary truths.

So maybe it’s saying that the “sublime consonant” can’t be debunked?! I say, therefore it is? Sort of. It seems like the act of naming is a weighty business. Pinker goes on:

Perhaps it is not surprising that people in so many cultures think words have magical powers…, or that one of the gospels should begin, “In the beginning was the Word, and the Word was with God, and the Word was God.”

And that’s exactly what was going on with Osho, although in this case he’s talking about the other side of the coin. Words create meaning, and in some cases that’s undesirable — as in when you can’t express what you’re talking about. It restricts and binds in ways that have profound implications:

Because he has never talked about God, many think Buddha is an atheist — he is not. He has not talked about God because there is no way to talk about God. All talk about God is nonsense. Whatsoever you can say about God is going to be false. It is something that cannot be said.

Other seers also say that nothing can be said about God, but they do say this much, that nothing can be said about God. Buddha is really logical. He will not say even this, because he says, “Even to say that nothing can be said about God, you have said something. If you say ‘God cannot be defined,’ you have defined him in a negative way — that he cannot be defined…’”

Ludwig Wittgenstein, one of the greatest thinkers of this age, one of the greatest of all the ages also, has said, “That which cannot be said must not be said. That which cannot be said, one must be silent thereof.”

Finally, both sides of the issue — that words describe reality while at the same time creating it — are explored by Bohm in his concept of “enfoldment”.

… each perception of a new meaning by human beings actually changes the overall reality in which we live and have our existence — sometimes in a far-reaching way. This implies that this reality is never complete.

Does it converge? You can be the judge of whether I should have torn out a page or two, rolled’em up and smoked’em. But even if it’s all a mirage, we’ve still got those -ngr answers…

Expert Testimony

Since the participants in the Beijing-R discussion were gracious enough to allow reprinting, I’ll try to keep the editorial work to a minimum and simply reorganize the conversation. Comments come from John Wells, Nigel Greenwood, Kwan-Hin Cheung, and Victor Mair and here’s what they had to say about the -ngr and about Mandarin rhoticization in general.

How would you describe -ngr, phonetically?
The starting hypothesis was to characterize it as a “nasalized retroflex approximant.” But after some conversation, the conclusion seems to be that the effect is to produce a nasalized, retroflexed vowel.

Wells: The obvious way to write tangr would be [tʰɑŋɻ], if that is how it is pronounced (i.e. with a sequence of final consonants)… However it may be that the r-effect is merged with the velar nasal to make a single segment. If that is so, perhaps we could write [tʰɑŋ˞], using the IPA rhotacizion diacritic. Or is the result actually a nasalized retroflex approximant? If so, then [tʰɑɚ̃]… A rhotacized ŋ would have contact between the back of the tongue and the velum, with the velum lowered (= the specification for ŋ) plus retroflexion of the tongue tip (which would not have much effect on the resultant sound, since there is no air passing through the mouth cavity). A nasalized retroflex approximant would be a sound something like like AmE ɚ in summer or dinner, i.e. with the soft palate lowered during the vowel because of allophonic assimilation of nasality. I should think this is more likely to be what you’re hearing.

Greenwood: The most concise description I know of this pronunciation comes in YR Chao’s Grammar of Spoken Chinese. In the section “Morphophonemics of the Retroflex Ending” on p46, he has this to say (note that in his GR transcription the retroflex -r is written -l):

With the back nasal ending -ng, a compromise is followed by having the preceding vowel both nazalized and retroflexed, thus: sheng + -l ->shengl = [ʂə̃˞]

Do linguists consider -ngr to be distinct from -nr?
This question came up in the Woods, Pears & Jingle Bells guest post back in February. The short answer is yes. Mair quotes a straightforward er-ization rule from memory out of the same Chao grammar…

Mair: What I remember (albeit vaguely) from those far away days is basically this: (V stands for the YUNMU or vowel of the syllable)

V + nr sounds just like V + r
V + ngr resiults in nasalized V + r

And Cheung sends a page from the Encyclopaedia Sinica that comes to the same conclusion. (Note that those are not tone marks over the vowels, but tildas indicating rhoticization)

Is the -ngr “exotic” — i.e. occurring in few other languages?

Wells: It’s an allophone of AmE syllabic r. I think you would certainly expect to get it in a phrase such as minor name, between the two n’s. Apart from Mandarin and American English, an r-coloured mid central vowel is extremely rare in the languages of the world. It just happens to occur in two of the most widely spoken ones. I’m sure that this is true of the nasalized equivalent, too.

Greenwood: I don’t know of any other language that has precisely this combination of retroflex and nasalization.

Any tips for Zhonglish speakers on learning -ngr?

Greenwood: If you had to get across the pronunciation of “tangr” to a non-native speaker of Chinese, I suppose you could suggest trying to say a rhotic (eg General American) “tar” while simultaneously pinching your nose. The trick is then to let go of your nose without losing the twang!

Mair: I have tried to pronounce -ngr just the way it is spelled, but it seems almost physically impossible to really make that velar come out at the same time as one is nasalizing and rhotacizing. Moreover, when I hear Chinese speakers talking, the nazalization of the vowel that results from V + ngr is really quite conspicuous, and the rhotacization is there, but the velar quality is absent, whereas, when I hear Chinese speakers pronounce V + nr, the vowel is not nasalized, but the rhotacization is obvious.

The Ultimate Beijinger

All this talk starts to get uncomfortable — masochistic, nearly. The Beijing-R fetish isn’t meant to be riled up like this; it’s dangerous territory. The double-/z/ opulence of the fricatives in “nasalized” hum like the Siren’s call. Běijīnghuà is beckoning! You catch yourself unconsciously hanging on the initial rhotic of “retroflex”, hoping against logic it might offer a little more than just description…

Never fear. The nasalized retroflexed vowel is close at hand. No talk. All sound. The real thing. Out of Beijing Sounds cold storage we have rescued a recording from Beijing TV. At the time, no one could see what purpose it might serve, but its timely rediscovery will serve to justify the business expense of such lavish archiving facilities.

The recording suffices to prove that not every TV personality has had their Beijing Dialect harmonized to exacting pǔtōnghuà (普通话, i.e. standard Mandarin dialect) standards. Staff analysts claim to have identified a dozen Beijingisms in just the intro (here’s the whole thing, not Firefox friendly). I’ll leave it to the reader to see if they deserve their jobs and simply point out, in the third line below, the ability to turn what might be three syllables in some world ( dàn yǒu rénr = 但有人儿) into one extended rhotic conglomeration. Beautiful.

[Go to website or bottom of this post to listen to audio]

说的一个字儿， 想必大伙儿都很感兴趣
Shuōde yīgè zìr, xiǎngbì dàhuǒr dōu hěn gǎn xìng qu
This one word, presumably everyone has an interest in

就是钱呢。咱北京和这钱有关系的胡同儿，也有这么几条
Jiùshìr qián ne, zán Běijīng hé zhèi qián yǒu guānxi de hútòngr, yě yǒu zhème jǐ tiáo.
It’s money. And in our Beijing there are a few neighborhoods that have a connection with this money.

但有人儿说了，这胡同儿以前肯定是印钞厂
Dàn yǒu rénr shuō le, zhèi hútòngr, yǐqián kěndìng shì yìnchāochǎng
But there are people who have said that these neighborhoods earlier were definitely money printers.

您说的不全对。所以今儿各我跟你聊了北京和钱有关系的胡同儿。
Nín shuō de, bù quán dùi. Suǒyǐ jīnr ge wǒ gēn nǐ liáo le Běijīng hé qián yǒu guānxi de hútòngr.
You said it wasn’t completely right. So today I’m going to chat with you about Beijing’s neighborhoods with money connections

Everyone knows that literacy in Mandarin means hour after brutal hour of memorizing and practicing a script whose design clearly shows the influence of sadistic genius. Here are a couple of favorites from the torture rack: two pairs of characters that have absolutely no connection except that they just might possibly, to the benighted, appear to be vaguely similar in form to one another.

衣农， 日曰

Well, at least we think we know what Mandarin literacy means. But do we really? Could you possibly get away with achieving literacy through something less than masochism? A couple of prominent writers have recently hinted that perhaps it does not have to be as bad as all that (for the foreign learner, we’re talking about right now).

On pinyin.info Mark Swofford, who has been known to use the M-word (masochism) in relation to character-learning, has compiled a list of readings in Pinyin. If you couple those with a pinyin-based dictionary like the ABC (which is contained in the Pleco dictionary by the way, one of my favorite Mandarin tools — I keep meaning to rave about it at length), you could find yourself reading real Mandarin, including literary works, in blissful ignorance of Chinese characters! By the way, the post also references a good book on pinyin orthographic conventions along with a sample chapter! This is so good that I should devote more space to it in another post.

On Language Log, Victor Mair also invokes the M-word in connection with characters, but continues on an optimistic note with the news that a just-launched newspaper has characters paired with pinyin for pronunciation. For Mandarin teachers he hopes this could be a step in the right direction, away from “brute force” early-stage memorization of characters and towards something more like “benign pedagogical methods” that emphasize acquisition of the language up front and only later introduce the script.

So there’s the gist of it: you can now, in some enlightened places, study Mandarin as a foreigner without immediately being crushed under the weight of a script that has gone thousands of years without linguistically significant reform. Even better, you can access resources that allow you to read Mandarin, picking up pronunciation and vocabulary, without the added burden of memorizing a few thousand characters. But can you really get away with ignoring the ubiquitous hànzì? Probably not. Hence more time spent some days with the Pleco dictionary than with the spousal unit. And I’d have to agree with John at Sinosplice that Mair’s article seems to exaggerate the joys of using pinyin to learn characters. It’s still damn hard.

Thanks, Mr. Zhou

Even though you can’t avoid the pain, you should still be grateful to this Shànghǎirén (上海人, person from Shanghai) who has adopted Beijing as his home and is well known as one of the key players in the creation of Pinyin. We’ve seen a bit of Zhōu Yǒuguāng (previous post here and video here), but it would be nice to know a little more about what he’s been thinking. What does he think the point of Pinyin is, anyway? I was pleasantly surprised to turn up, after a bit of browsing, a video that not only is recent, but also offers some interesting history and substantial discussion.

Here’s all 43 minutes, and a link if the embedding doesn’t work:

That’s a lot of video, so just a few observations below: on Hanyu Pinyin as a two syllable term; Pinyin and the topolects; how learning characters could be worse.

1. The two syllables of “Hanyu Pinyin”

Sometimes even when you know exactly what someone’s going to say, it’s hard to understand it. Earlier discussions have covered elliptical Beijing phonetics, where /sh/ or /xi / or /zh/ get dropped from casual speech. The first speaker from the audience has the full range of that — even check out the nearly-dropped /ch/ in fēicháng in the dialog below.

But turning the four syllables of hànyǔ pīnyīn into two is a new one for me. Try this on for size:

[Go to website or bottom of this post to listen to audio]

Host:
Lánsède zhèige cíbǎnshang guàde shì shénme?
蓝色的这个词板上挂的是什么？

Beijinger:
yīgè, sāngè fēicháng kěpàde zì
一个，三个非常非常可怕的字。

Host:
这是英语吗？

Beijinger:
Zhè yīnggāi shì hànyǔ pīnyīn
这应该是汉语拼音…

Now pinpointing the beginning/end of a syllable is tricky business for anyone (if someone can Praat this, I’d post it in a nanosec), so feel free to argue that it’s still four. But you have to admit that by not closing off the N in either “hanyu” or “pinyin”, as is common among Beijingers at least, he comes pretty close to giving us just a nasalized vowel sound in the first word and an extended /i/ in the second.

As long as we’re taking this side tour away from Pinyin, it’s worth one more quote from our Beijinger and the host to hear a snippet of discussion on housing prices as they relate to salaries. It’s a topic of endless conversation in Beijing. Folks know price-per-square-meter figures the way Americans know gas prices. Since it’s a safe topic among strangers, small talk about housing prices is as pervasive (and mournful in tone) as small talk about the weather in Minnesota.

[Go to website or bottom of this post to listen to audio]

Host:
Ei wǒ tīngguo yīgè zhuānjiā gēn wǒ shuōguo,
唉我听过一个专家根我说过
Hey, I’ve heard an expert tell me

tā shuō [shuō] zhèige hélǐde fángjià
他说这个合理的房价
he said these housing prices would be rational

shì zhèige chéngshì jūmínde yuèpíngjūn gōngzī
是这个城市居民的月平均工资
if a city resident’s monthly salary

yuèpíngjūn gōngzī néng mǎi yīpíngmǐ
月平均工资能买一平米
monthly salary could buy one square meter

tā rènwéi zhèige fángjià jiùshì [jiùshì] héshì le
他认为这个房价就是合适了
[then] he figures these housing prices are about right.

Beijinger:
Yě chàbuduō
也差不多
Sounds about right.

Host:
Nǐ xiànzài yuè gōngzī yě sānwàn duō ma, chàbuduō?
你现在约工资也三万多吗，差不多？
Your monthly salary is now also about 30k RMB more or less, right? [i.e. about $4000/mo -- outrageously high for an average Chinese]

Beijinger:
Mǒmo yīgè língr chéng ma?
抹抹一个零儿成吗？
Could we wipe out a zero?

Any opinions on the character for the host’s first utterance, “ei”? The ABC dictionary has that character (唉) as “ai” — an interjection.

Without further information, this dialog is pretty strong evidence the host is a Beijinger as well. He’s close to dropping a couple of SHs, e.g. especially in jiùshì (就是). The respondent then does a classic Beijing-R-ification with língr, which connotes, as others have noted about érhuà (儿化) a diminutive aspect.

2. Pinyin might hold hope for the topolects

After the show’s host brings up the subject of Pinyin with Zhou, he asks the audience what it’s good for. The first answer is “foreigners learning Chinese” — fine as far as that goes. I’m certainly a believer. The second is something about pinyin giving Chinese the best of both worlds: an alphabetic system and the “ideographs” (象形 = xiàngxíng). Not sure what to make of that.

But the third audience member brings up the idea that pinyin can be used to write anything, especially the fāngyán (topolects, to use Victor Mair’s word) i.e. the Wu language, Cantonese, etc. Now leaving aside the fact that Pinyin is getting conflated with the more general idea of phoneme-based writing, the assertion is still interesting. It shows that this (presumed) layman is quite aware that the topolects are not generally written using characters, that what gets written is, in fact, just pǔtōnghuà i.e. standard Mandarin. The topolects are, with some limited exceptions, scriptless. This is a shame, as every language deserves a good script.

The suggestion that all the topolects should be romanized has been made in China since at least the 1890s, as discussed in DeFrancis’s superb book, Chinese Language: Fact and Fantasy (Pinyin.info has a sample chapter). From page 242:

Perhaps the central issue [among the first reformers in the 1890s] was just what kind of Chinese speech should be represented by the new alphabets. One group wanted to confine their use to representing Mandarin. Another group argued that literacy even in alphabetic Mandarin was too difficult for some of the illiterate masses and insisted on creating separate phonetic schemes for the various regionalects and relegating use of the Mandarin-based script to situations in which a national orthography was essential.

Clearly the latter has prevailed, to the extent that Pinyin is adopted as a script at all. If I had a vote, I’d put it in with the former group as a way of promoting full literacy.

3. Learning characters could be even worse

And then there is Zhou’s answer. What’s Pinyin good for? He comments on how it allows you to figure out the pronunciation of characters without the benefit of having a teacher around to tell you. Sounds like a duh, but had you ever thought about how bad things could be? Here’s the system he describes

[Go to website or bottom of this post to listen to audio]

Zhou:
Wǒ xiǎode shíhòu dúshū
我小的时候读书
When I was young and reading books,

yīgè zì bùrènshi jiù yào wèn lǎoshī
一个字不认识就要问老师
if I didn’t recognize a character I had to ask the teacher

chá zìdiǎn bùxíng
查字典不行
It wouldn’t work to look in the dictionary

Host:
shīfu dài túdi
师傅带徒弟
Master leads apprentice

Zhou:
yīnwèi zìdiǎn shàngmian yòng fǎnqiède fāngfǎ lái zhùyīn
因为字典上面用反切的方法来注音
Because the dictionary uses this method of joining two characters to indicate pronunciation

Yào rèn liǎngge zì lái zhù yīgè zì xiǎoháizi bùnéng yòng
要认两个字来注一个字小孩子不能用
If you have to recognize two characters to pronounce one character, children won’t be able to use it

Ouch, no kidding. I knew nothing about fǎnqiède fāngfǎ (反切的方法) before this. A quick Baidu search turns up this explanation (in Mandarin, in characters) which links to the graphic below, which you’ll notice just can’t resist adding in a few Pinyin consonants just to help the explanation. Try that for a while, then thank your lucky stars you’ve got Pinyin.