How to pronounce 小姐?

[querido]

Pthcs (pinyin #21) gives this as the pinyin: Xiǎo·jiě (With that dot. Why is that?)

Hear the attachment.

21.mp3

[Daan]

Thanks for posting that recording. As you would expect, that speaker is applying the tone sandhi rule and pronouncing the word as xiáojiě, so 23, not 33.

To your examples, the rules for tone sandhi in Mandarin are very simple: 3rd tone -> 2nd tone before another 3rd tone as part of the same word, 不 -> 2nd tone before another 4th tone. I'm probably missing some other examples as well, but my point is that you just learn the rules, and that's how you know how it is pronounced.

Well, there's a few slightly more confusing tone sandhi rules, but they're facultative and almost never taught to learners. Instead, they're acquired naturally as you start speaking more fluently. For example, there's the rule that when a 1st or 2nd tone is followed by a 2nd tone and then by a non-neutral tone, the second 2nd tone will be pronounced in the 1st tone. But on the whole, I agree wholeheartedly that the basic rules aren't that difficult, and can easily be applied by learners. (By the way, you'd also expect to see 33->23 tone sandhi if you have a couple of monosyllabic words in a row, where the syllables are all originally in the 3rd tone. I can't come up with an example at the moment, but it doesn't only apply within words.)

Note that pinyin does show tone qingsheng (5th tone) (did I get that term correct?), at least MDBG and the HSK word lists I have do. [Although, weirdly enough, the MDBG entry for 老虎 shows 2 third tones. Error in MDBG?]

Pīnyīn does indicate neutral tones (or qīngshēng/5th tone, indeed), but that's partially responsible for the problems here. If we indicate neutral tones but no tone sandhi, you get lǎohu. You can't tell by looking at lǎohu alone that the last syllable in that word was originally pronounced as hǔ in isolation, and that you therefore need to pronounce the first syllable as láo rather than lǎo.

The creators of Hànyǔ Pīnyīn could have decided not to indicate neutral tones, but that would have caused yet more problems (because there are a lot of words with neutral tones) and in this case would still not have reflected the accurate pronunciation, giving láohǔ instead after applying the tone sandhi rules. As it stands, there's only a few words affected by not indicating tone sandhi while indicating neutral tones, but the entire problem would not have existed if tone sandhi, at least within words, were indicated in Hànyǔ pīnyīn. I hope you see why I'm always complaining about this

[Shi Tong]

bù kěyì is always the way I hear bu4 ke3yi3.. despite the fact I know it's supposed to be the latter.. not only that but I always pronounce it the first way bù kěyì and I've never been corrected ;)

Probably people being too polite... and yes- I agree it's STRANGE.. it's also strange in all the other examples I sited.

By the way, I think it may come out as bù kěyì because of the usual stress that people put in their voices when speaking to people saying "dont do that".. maybe the anger of "dont do that" makes the yi3 into an yi4.

I do agree that not knowing how to pronounce certain things where second zi's in words go neautral is really annoying- this is something which happens less in Taiwan, so I dont get as much bother on the pronunciation of neautral tone as I imagine some mainland learners get.

EDIT- BTW- yes, sounds exactly as I expected it would- xiao2jie3... indicating that you pronounce both the tones (jie doesn't turn into 0) and the xiao which is usually xiao3 becomes xiao2.

[xianhua]

It also depends on where the 小姐 appears in the sentence. If it's in the middle, then 小姐 becomes 2/3, with the second one taking on a partial third tone only. If it's at the end of the sentence, or standing alone, you may hear 2/3 - with the 姐 taking on an exaggerated and drawn out full-third tone from the impatient diner.

[jbradfor]

I hope you see why I'm always complaining about this

Actually, no

I agree that having the tone markers the way you describe would make reading pinyin easier. But don't you think it would make writing correct pinyin harder, due to more rules? Given that aspect, and how simple the rules are, it doesn't seem that big a deal to me either way.

But I'm surprised you didn't say anything when I called pinyin only a system of "transliteration". Given that pinyin does indicate qingsheng, which is related to spoken only, upon reflection doesn't that mean that pinyin does more than just transliteration and also includes some aspects of transcription as well?

[xianhua]

As for 老虎, 老 and 虎 are both third-tone in isolation, so we have 3/3, but 2/3 with Sandhi. Why would 3/3 be considered incorrect?

One negative knock-on effect of changing pinyin tones based on sandhi would be to confuse new students into thinking that the word always retained the given tone. We'd then have students thinking 老 was always a second tone, and addressing their 老师 in a 2/1 combination which would just be awful.

[renzhe]

For what it's worth, the 新世纪汉英大词典 also states xiao3jie3 (so 33 -> 23)

It's the more conservative pronunciation, but the qingsheng variant is more common in colloquial speech, at least according to my experience. And then it's 33 -> 23 -> 20.

[mirgcire]

you didn't say anything when I called pinyin only a system of "transliteration".

I may not be the "you" that you are addressing, but it is hard not to object to your use of the word "only".

Just for the record, according to Wikipedia, "Transliteration is the practice of converting a text from one writing system into another in a systematic way."

Correct me if I am wrong, but transliteration is a relatively rare use of pinyin, only used in situations where you simply don't have the Chinese characters available, like sending an email or texting.

Pinyin was developed as pedagogical tool to provide a phonetic representation of mandarin. It's primary use is to allow Chinese school kids a way to transition from the speaking to writing. And it's second most common use is probably teaching foreigners, either as a transitional tool or as a way to avoid learning the writing system.

And for what it's worth, I do agree that dictionaries should apply tone change rules and represent words as they are actually pronounced. And it seems particularly inane for them to take the liberty of representing 小姐 as 3-0 but representing the bulk of other 3-3 words (for which tone change rules apply) as 3-3.

[jbradfor]

@xianhua

As for 老虎, 老 and 虎 are both third-tone in isolation, so we have 3/3, but 2/3 with Sandhi. Why would 3/3 be considered incorrect?

Because Daan was stating (and I think this is correct, although I can't find a reference), that 老虎 is qingsheng, that is, the tone on the last characters goes to the neutral (aka 5th) tone. Since MDBG states that they have the pinyin account for qingsheng but not sandhi, then the tones on 老虎 should be listed as 3/5.

[Altair]

Here is what nciku.com has for 小姐, 不可以, and 老虎 (Press "Listen" to hear the pronunciation).

Pthcs (pinyin #21) gives this as the pinyin: Xiǎo·jiě (With that dot. Why is that?)

I think the dot is used in learning materials to show that the following syllable is 轻声.

bù kěyì is always the way I hear bu4 ke3yi3.. despite the fact I know it's supposed to be the latter.. not only that but I always pronounce it the first way bù kěyì and I've never been corrected ;)

Is your pronunciation really different from the recording above? How do you pronounce: 可以不可以? With a different pronunciation of 以 for each occurrence? Are you sure you aren't mistaking the fall in pitch on this syllable with a fourth tone?

One negative knock-on effect of changing pinyin tones based on sandhi would be to confuse new students into thinking that the word always retained the given tone. We'd then have students thinking 老 was always a second tone, and addressing their 老师 in a 2/1 combination which would just be awful.

After an initial annoyance at the lack of showing the sandhi, this is the conclusion I reached. I preferred knowing that I could trust all the second tone marks to being reminded of sandhi that I could learn through rules.

And for what it's worth, I do agree that dictionaries should apply tone change rules and represent words as they are actually pronounced. And it seems particularly inane for them to take the liberty of representing 小姐 as 3-0 but representing the bulk of other 3-3 words (for which tone change rules apply) as 3-3.

I think the reason for the preference for indicating 轻声, rather than sandhi, is that 轻声 can occasionally differentiate meanings in words spelled with identical characters (e.g., 精神 jīngshén (mind) and jīngshen (energy)).

[Mugi]

Or they account for tone neutralization, but not for tone sandhi - which is what mine do, if I recall correctly.

In which case such dictionaries fail in their responsibility to systematically indicate correct pronuciation - it's bizarre to indicate a secondary potential phonological change, but not a primary compulsory change. Unless there is some other orthographic inidcation, it would seem that such dictionaries can't distinguish between a 30 combination that does undergo tone sandhi and one that does not.

[Daan]

Mugi's hit the nail on the head. That's exactly what I've been trying to explain

Let's take a look at the word for 'tiger' again, assuming that láohu and láohǔ are both accepted pronunciations in different varieties of Mandarin after applying tone sandhi and neutralisation rules, which I don't believe anyone has questioned.

lǎohǔ will lead learners familiar with the tone sandhi rules to pronounce this word as láohǔ, which is an accepted pronunciation in southern Mandarin, but does not tell them hǔ can also be pronounced in the neutral tone. Yet láohu is the standard pronunciation in Modern Standard Mandarin (MSM). This romanisation, then, indicates neither tone sandhi nor neutralisation, and as a result requires learners to know the tone sandhi rules, as well as leaving them in the dark about the standard pronunciation in which neutralisation occurs.

láohǔ works for learners unfamiliar with tone sandhi rules, but does not tell them the second syllable can also be neutralised.

lǎohu is arguably the most 'dangerous' romanisation. The neutral-tone syllable was originally a third tone, and thus affects the pronunciation of the first syllable because of the 33->23 tone sandhi rule. A learner who did not know this could not possibly be expected to come up with either láohu or láohǔ. A teacher would need to explicitly explain this to his students, or his students would have to be in the habit of looking up all the characters they encounter in a dictionary to check their original tones. But what if students aren't learning characters at all (yet), focussing instead on speaking and listening only?

láohu, finally, gets learners straight to where they need to be: the standard pronunciation of the word for 'tiger'. It does this by indicating both tone sandhi and neutralisation.

Which romanisation to choose, then? I think the best idea would be to use both láohu and láohǔ, explaining that they're identical in meaning, but used in different varieties of Mandarin. Looking at láohu and láohǔ, learners unfamiliar with tone sandhi and neutralisation rules would be able to pronounce both forms correctly immediately, with the first romanisation obviously referring to the MSM pronunciation, and the second one to the pronunciation more commonly heard in more conservative Mandarin.

Just for the record, I'm not saying learners should not learn the tone sandhi rules. They're important and should be taught to learners early on in their studies. But the creators of Hànyǔ pīnyīn have simply failed to recognise how their decision not to indicate tone sandhi, while at the same time indicating neutralisation, causes problems where a neutralised syllable affects the pronunciation of a previous syllable because of the 33->23 tone sandhi rule, since it cannot be clear from the Hànyǔ pīnyīn that the neutralised syllable was originally pronounced in the 3rd tone.

[anonymoose]

lǎohǔ will lead learners familiar with the tone sandhi rules to pronounce this word as láohǔ, which is an accepted pronunciation in southern Mandarin, but does not tell them hǔ can also be pronounced in the neutral tone. Yet láohu is the standard pronunciation in Modern Standard Mandarin (MSM).

On what basis are you claiming that láohu is the standard pronunciation in Modern Standard Mandarin?

[Daan]

Isn't it? That's what I was taught, and I've never been corrected or misunderstood when pronouncing the word as láohu. So I think I'm claiming that on the basis that that's what people say, if I'm not very much mistaken. And I may very well be mistaken, but even if that were not the standard pronunciation, the case against not indicating tone sandhi would still hold because of 小姐, for example.

[anonymoose]

And what makes you think xiáojie is the standard pronunciation?

Both the 现代汉语词典 and "A New Century Chinese-English Dictionary" list 老虎 and 小姐 as lăohŭ and xiăojiĕ respectively. It can't just be a convention of the dictionary to list all characters with their original tones, since other words that are supposed to have a neutral tone are listed as such, for example qīngchu (清楚) or chōuti (抽屉), and I doubt those dictionaries would list "southern mandarin" rather than "standard mandarin".

[Mugi]

Just for reference...

Digital version of 中日辞典 (小学館) has the following:

. 小姐 xiǎojie ... 発音：実際はxiáojieと発音する. [Pronunciation: Actually pronounced as xiáojie.]

. 老虎 lǎohǔ ... 発音：láohuとも発音する. [Pronunciation: Also pronounced as láohu.]

The sound files for the above words are clearly "xiáojie" and "láohǔ" respectively.

Digital verson of 中日大辞典 (大修館書店):

. 小姐 xiǎo•jiě

. 老虎 lǎo:hǔ

I can't decide whether the accompanying sound file pronounces 小姐 as "xiáojiě" or "xiáojie", but 老虎 is clearly pronounced as "láohu".

I suspect that the single dot between syllables indicates that the final syllable is actually pronounced in the neutral tone, and that the colon means the final syllable can be pronounced in the original tone or neutral tone.

汉语方言词汇(第二版) (语文出版社) has the following for 老虎:

. 词目: 老虎 lau3→2 xu3 . 北京: 老虎 lau3→2 xu3, lau3→2 xu5

*Romanization is in China-modified IPA; tone bars have been changed to numerical representation by me.

西安 also has lau3→2 xu5 (though interestingly, not lau3→2 xu3).

To me, both "láohǔ" and "láohu" sound natural, but I probably tend to say the latter.

[jbradfor]

@Daan, you now have me halfway convinced. HALFWAY

On one hand, I agree that the case you present, two nominally third tone characters where the second character has qingsheng, is quite an issue. [Whether 老虎 is an example of this or not is immaterial to this discussion, as long as there is one word that does this, it is a problem.] I can't think of any (reasonable) solution besides the one you suggest, i.e. indicate sandhi in pinyin.

On the other hand, I'm not sure it is worth adding more rules to pinyin just for this one case. I mean, people can't even get the rule for the apostrophe correct, and that's really simple.

So it comes down to slightly easier reading pinyin and removing an ambiguous corner case, versus much easier writing pinyin.

Will ponder it more. Maybe next week I'll be more convinced.

[anonymoose]

Unfortunately there's no such thing as a perfect system. Standard pinyin does pretty well to represent the tones, except in this rare case as pointed out. I don't think it's worth getting one's knickers in a twist over though. I mean, apart from tones, multisyllabic words also have stressed and unstressed syllables, and noone's complaining about that not being marked in the dictionary.

And incidentally, I think a lot of cases where there is disagreement over whether a final syllable should have a tone or be neutral is actually because it's an unstressed syllable, and to some extent these tend towards being neutral even though they nominally still have a tone. (And there's also quite a lot of variation between different speakers.)

[Shi Tong]

What I think they should do is mark the things twice- once for how each word is traditionally pronounced (like lau3 and hu3) and then another pronunciation key for how you say these together (as lau2hu3 or lau2hu).

Personally, I know what I've always heard is lau2hu3 and xiao2jie3- obviously both with sandhi.. but the confusions of qingsheng always play havok with some of these combinations.

As for bukeyi-- Of course, your site is showing that it's bu4 ke3yi3, but that's because it shows each sound individually and takes absolutely no sandhi or colloquialism into account.

Of course, I've heard bu4ke2yi3 before (with the sandhi on the ke3), but it's only said this way when the sayer isn't frustrated or anything, and I've never ever said 可以不可以.. only 可不可以.. either way I still pronounce these using a fourth tone for the last yi4.

It's also not like the yi at the end has to be a really hard fouth- it's more of a soft fouth or half fourth. I've always said it like that and nobody has ever blinked an eyelid.

That said, I still also know it's incorrect, it's just one of those odd things I've noticed, much like the confusion of xiao3jie3 and lao3hu3.

[carlo]

I suspect that the single dot between syllables indicates that the final syllable is actually pronounced in the neutral tone, and that the colon means the final syllable can be pronounced in the original tone or neutral tone.

Something like this is found in most dictionaries. Take 现代汉语词典, which is a sort of putonghua standard. Here's what it says in the 凡例:

条目中的轻声字，注音不标调号，但在注音前加圆点，如便当 biàn●dang

一般轻度、间或重读的字，注音上标调号，注音前再加圆点，如 因为 注作yīn●wèi

xiaojie belongs to this latter case. As to how "standard" is the standard, we could argue for years. In this case it's again a compromise between Beijing pronunciation and national usage.

"bu4ke3yi4" may help you pronounce this correctly, but as a transliteration it's wrong in more than one way. Check: 1. Say 不可思议, then take out the third character, do they sound the same? 2. Is the starting point of the "yi" sound higher or lower in pitch than the ending of the preceding vowel "e"? 3. Is the vowel in "ke" rising or falling in pitch (ie a second tone or a half third?)

AFAIK dictionaries do not distinguish between the two types of "33" combinations, which is a problem.