neuter-tone syllables and tone-shift

[Alchemist]

I have some doubts about tone shifts when there's a neuter tone around. Let me explain...

I know that 《一个》 is pronounced "yi2 ge", so it looks like 一yi1 still "senses" the original fourth-tone pronunciation of 个, even though 个 is actually pronounced with a neuter tone. All of the books that I have seem to agree on this pronunciation.

Now let's take a look at 奶奶 and at 小姐.

I have a book that says that 奶奶 is pronounced "nai3 nai" (where the first "nai3" is actually a half third-tone). So here it looks like the second 奶，after losing his original third-tone pronunciation, doesn't force the third-to-second tone shift on the preceding 奶.

I have another book that says that 小姐 is pronounced "xiao2 jie". So, here it looks like 姐 does force the tone shift on the word 小, even though it lost its original third tone.

Actually, I'm puzzled...

How do you pronounce 一个，奶奶 and 小姐?

Can anyone please enlighten me about the way neuter-tone syllables affect the pronunciation of the preceding syllable?

[Altair]

The details of what you have posted are correct. The compounds are pronounced as you have indicated. Your confusion comes from the fact that grammar books generally ignore that linguistic rules are often more complex than they seem. They simplify things to make the early learning easier, even if it results in greater confusion later.

I once read a linguistic paper on this subject. What the author argued was that the differing treatment of third-tone syllables before neutral-tone syllables stems from the fact that various grammatical rules of Chinese (and of all languages) are not applied at the same level of theoretical structure. The output of the rules ultimately appears together, but the rules themselves must be applied in a preset sequence. If you change the sequence, you change the outcome.

The paper talks about “tone-deletion” rules that create neutral-tone syllables. The details of these rules are quite complex and probably not fully understood; nevertheless, it seems that all neutral-tone syllables are not created equal. For the issue at end, it makes sense to think of reduplicate nouns and words ending in the suffix 子 (zi) in one group (Group A). (E.g., 奶奶 nainai, 姐姐 jiejie, 椅子 yizi, 种子 zhongzi, and 脸子 lianzi. 耳朵 “erduo” really belongs here, as well.) Reduplicated verbs, reduplicated adjectives, or combinations of verbs or adjectives with nouns can be thought of as comprising another group (Group B). (E.g., 点点 diandian, 找找 zhaozhao, and 小姐 xiaojie).

The Group A words can be thought of as primarily lexical. The original tone of the second syllable is deleted by a rule that seems to act at the lexical level of the language. Components of the Group B words seem to retain a greater sense of independence or clearly result from an active grammatical process in the language. For Group B words, the tone deletion rule seems to operate, if at all, on a more conscious level of grammar. Sometimes the tone is deleted, and sometimes not.

Apart from these two rules is the rule that changes third-tone syllables into second-tone syllables. To put these rules together, they must be applied in the following order: (1) delete the tone of the second syllable in words in Group A, (2) change third-tones to second tones whenever they precede another third tone (This rule is actually a great deal more complex than this.), and (3) reduce or delete completely the tone of the second syllable in words in Group B. The result of this sequence is that the Rule 2 has nothing to work on in words that are in Group A, and the first syllable will retain its third tone pronunciation. Rule 2 will, however, apply to all words in Group B, but the tone deletion rule may sometimes eliminate the evidence of the need of it.

An apparent exception to this rule is the pronunciation of the word 法子fazi. As far as I can recall, fa2 zi5 is the usual pronunciation, despite the fact that fa3 zi5 is the pronunciation one would expect. I think this latter pronunciation may, however, be acceptable to some speakers. Perhaps other can comment on this.

The pronunciation of 法子 as kong2 zi3 would also seem to be an exception, except that here, the tone of zi3 is not deleted and the pronunciation of “kong3” can undergo its regular change into “kong2.”

In sequences like 一个 yi2 ge, the change in “yi” can also be explained by applying the rules in sequence. In Mandarin, all measure words are normally unstressed when they follow a numeral. In the sequence, yi2 ge4, the “ge4” will not normally be stressed. While some measure words have a significant semantic content, “ge4” has almost none. Mandarin words that have no semantic content tend to lose their tones, particularly if they are not stressed. This tone deletion rule is applied after the fourth tone of “ge” works its magic on “yi1” and changes it to “yi2.” If “ge4” is stressed in speech or is pronounced with exaggerated care, its tone is retained, producing the sequence “yi2 ge4.” This is the pronunciation given on some cheap language materials that strive to teach only a few dozen phrases or give an introduction to Chinese. I think the pronunciation “ge4” is also normal in other sequences, such as “ge4 ge4” (“each and every”).

I hope this helps.

[HashiriKata]

This is a very interesting topic and the theory Altair cited has many valid points. I’m however inspired to making it a bit more accessible, more transparent to people without a background in linguistics. I think I’ve already got some basics for simpler rules but before that, I’ll need more data to be absolutely sure.

Please help by adding to the data base the third-tone compounds you know but are not already in the data base below, or giving the alternative tone pattern if you think some already in the data base have an alternative pattern or are simply not correct (Just give them in your post and I’ll edit this post & add them to the data base).

THE DATABASE

(Edit: the database has been moved to avoid confusion. Please see further on.)

[Alchemist]

Altair and HashiriKata, thank you for your replies. Your explanations were very helpful to me.

[HashiriKata]

I think I’ve already got some basics for simpler rules but before that, I’ll need more data to be absolutely sure.

My attempt to be the next Einstein seems to be completely ignored (or tone problems are not that important?).

2 weeks after my post above, I expect to get at least some feedback to say that the list above is not accurate or needs revising (I'm not a native speaker ). Since the list above will be the base for me to come up with some simple & useful rules, it has to be accurate and your critical eyes (or additional inputs) are still very much needed.

[Alchemist]

Since the list above will be the base for me to come up with some simple & useful rules, it has to be accurate and your critical eyes (or additional inputs) are still very much needed.

This won't be an additional input, but actually an additional question...

Which group does "哪里" belong to? Does 里 behave like 子, so that 哪里 belongs to group A?

[HashiriKata]

My instinct is that 哪里 belongs to Group B. (B), along with 小姐 xiao2jie5 and 法子 fa2zi5, so it should be 哪里 na2li5.

But this is precisely the kinds of things we need native & experienced speakers to confirm and help us.

[carlo]

I'm also a non-native speaker, but what are you saying seems ok to me.

普通话连读音变 (recommended, see: http://www.cp1897.com.hk/Search?Action=2&txtProName=%B4%B6%B3q%B8%DC%B3s%C5%AA%AD%B5%C5%DC&txtAuthors=&txtIsbn=&txtPublisher=) has a whole chapter on this issue. 法子 is also listed (if I remember correctly) as the only exception in the above categories.

[HashiriKata]

Thank you for your feedback, carlo.

I've looked at the link and it looks promising, the only problem is it may be difficult to get it from the UK. Is it a book plus the CD? Having tried it out, do you think it's a must for interested learners of Chinese?

Thanks also to Alchemist. Your input is now in the database. I think I'll be able to conclude this episode very soon.

Regards,

[HashiriKata]

So here is my promised attempt:

The rule: If the addition of the 2nd element does not contribute any meaning to the compound, then it does not influence the original tone of the 1st element.

As I’ve tried to condense the rule in order to make it easier to remember, please post if you feel something is unclear. The database is also re-arranged and reproduced here so that you can conveniently check it out. The rule is intended to cover all 3rd tone compounds involving the neutral tone.

Group A, tone pattern: 35 (2nd element not contributing meaning)

奶奶 nai3nai5, 姐姐 jie3jie5, 宝宝bao3bao5，婶婶shen3shen5, 姥姥 lao3lao5，嫂嫂 sao3sao5, 耳朵 er3duo5, 椅子 yi3zi5, 种子 zhong3zi5, 脸子 lian3zi5，鬼子gui3zi5，饺子jiao3zi5，本子ben3zi5, 脑子nao3zi5

Group B, tone pattern: 25 (2nd element contributes meaning to the compound)

(a) 点点 dian2dian5, 找找 zhao2zhao5, 走走 zou2zou5, 想想 xiang2xiang5 ( active, open-ended class)

(B) 小姐 xiao2jie5, , 哪里 na2li5

Group C, tone pattern: 23 (both elements with equal status & contributing meaning)

孔子 kong2zi3, 老子 lao2zi3

(No neutral tone is involved in Group C (cf. 水果 shui2guo3). Group C is included here only because 子 in this group may be confused with the 子 in Group A)

Exceptions: 法子 fa2zi5 (belonging to Group A but has tone pattern of Group B. This compound is also optionally pronounced as / fa3zi5/, in which case the exception disappears )

Note: If you are interested in the more basic principle of the change of the 3rd tone, please check out this thread: http://www.chinese-forums.com/viewtopic.php?t=2462

[carlo]

I've never ordered online from CP, I usually get friends in HK to buy books for me. I know there are Chinese bookshops in London (Gerard St?), would you be able to order through them?

This book is meant as a study aid to help Cantonese speakers improve their Mandarin pronunciation in connected speech, it's in traditional characters and with occasional references to Cantonese (but knowledge of Cantonese is not required). It has very good audio recordings on 2 CDs.

Not for beginners, but there are not many good books on Mandarin pronunciation for advanced students out there, and I think this is one of them.

[HashiriKata]

Thanks carlo!

I've just ordered it directly from the website. The postage to the UK is way above the cost of the book but I think it'll be worth it.

[carlo]

Apparently, a good half of the 'neutral tones' found in modern Mandarin Chinese originated in Beijing at the beginning of the Qing dynasty, when the Manchus were struggling to learn the Han language and were making the mistake that we as learners all make at some point: fall back onto the patterns of our native tongue. Stress in Manchu typically falls on the first syllable. So, while the 'native' neutral tones are grammatically justified (姐姐 is a reduplication), this 'Manchu pidgin' is kind of fuzzy. This is behind pronunciations such as 头发 fa0, 厉害 hai0, etc.

I found this fascinating snippet in a book I picked up yesterday and that promises to be a gripping read: 汉语语句韵律的语法功能, by 叶军 (an attempt to investigate the connections between prosody and grammar in modern Chinese).

[HashiriKata]

You're right. It's also very fuzzy when discussing the origin of tonal influence. Have a look at this related thread (where I have timely withdrawn ):

http://www.chinese-forums.com/viewtopic.php?t=3204

My rule above is intended merely as a practical guide when the learner is faced with a new compound, and I trust that once applied, the rule will be able to predict the correct/likely pronunciation.

[Altair]

Which group does "哪里" belong to? Does 里 behave like 子, so that 哪里 belongs to group A?

I am not a native speaker, but from my reading and listening, I am sure that 哪里 usually belongs to Group B, but can optionally be pronounced according to Group C in careful speech. 里 does not really lose or change its meaning and so cannot really be considered a purely lexical suffix like 子.

Since my earlier post, I have again reviewed the material in "Lexical Third Tone Sandhi and the Lexical Organization of Mandarin" by de Bao Xu in Chinese Phonology in Generative Grammar. As a result of what I read, I can add a few additional rules and clarifications:

Prefixes like 小 (xiao3) and 老 (lao3) will always undergo tone sandhi and change to second tone when the original tone of the following syllable is third tone: e.g., 小姐 (xiao2 jie5), 老虎 (lao2 hu5 or lao2 hu3), and 老鼠 (lao2 shu3 and lao2 shu?). When used as prefixes, I think they can belong to Group B or Group C, perhaps depending on the speaker, the circumstances, etc.. When used as adjectives, they always belong to Group C, for example, 小李 (Xiao2 Li3) used to address someone named Li.

Nouns created by reduplication belong to Group A, such as (小)本本 ((xiao2) ben3 ben5) (little booklet) and 寶寶 (bao3 bao5) (baby or "little treasure"). Kinship nouns like 姐姐 (jie3 jie5) (elder sister) and 奶奶 (nai3 nai5) (grandma) belong in this category as well.

Nouns being used as measure words and all adjectives and adverbs used in reduplications belong to Group B, such as 本本書 (ben2 ben5 shu1)(every book), 種種 (zhong2 zhong5)(every kind), 死死 (si2 si5)(tight, tightly), etc.

Verbs created by reduplication belong to Group A, such as 癢癢 (yang3 yang5)(itch) and 捻捻轉 兒 (nian3 nian5 zhuanr4)(top) and are different from already existing volitional verbs that are used in reduplications. These latter belong to Group B, such as 捻捻 (nian2 nian5) (give a little twist with the fingers), 想想 (xiang2 xiang5)(think a little).

One interesting thing to note is that characters 子, 本, and 捻 are treated differently depending on which pattern they belong to.

[Altair]

Someone pointed out some mistakes I made in my previous post in assigning the groups, although I think I had the correct tones. I have gone ahead and edited the post to correct it and to clarify some things.

Basically I think that Hashirakata's rule makes sense, but must be applied somewhat carefully. "Meaning" is a surprisingly subtle and specific term. For those who need to think in linguistic terms, I thought it made sense to add the additional material and did not bother to integrate it into the same format.

[HashiriKata]

Thank you for the additional material, Altair!

For those who need to think in linguistic terms, I thought it made sense to add the additional material and did not bother to integrate it into the same format.

For this purpose, I'd recommend 普通话连读音变 which carlo mentioned earlier. My order has arrived and I think it's an excellent buy for anyone interested in tone-shift in Chinese.