Mandarin, no voiced consonants?

[HSC]

According to the website 上海话入门， "Mandarin and Cantonese do not have voiced consonants."

From my understanding, voiced consonants produce vibration of the vocal cords, such as the word buzz.

When comparing voiced and unvoiced consonants, we usually see the comparison of the following consonant pairs:

Voiced/Voiceless: b/p, d/t, g/k

To me, words such as 百，狗 sound voiced. Even 排，开 seem to make the vocal cords vibrate.

Can anyone comment on this? I'm curious because I feel that my pronunciation may be off due to prior disregard for such details. I imagine that people who speak dialects like 上海话 must be sensitive to these differences even when listening to Mandarin Chinese.

[ala]

To me, words such as 百，狗 sound voiced. Even 排，开 seem to make the vocal cords vibrate.

It's the subsequent vowels that make these syllables vibrate (all vowels vibrate vocal cords), not the consonants. These consonants in Mandarin and Cantonese are voiceless.

Also, words like 百，狗 in Mandarin might sound kind of like English, because both are lenis, and English voicing isn't very strong.

Mandarin:

voiceless unaspirated --> lenis

voiceless aspirated --> fortis

Shanghainese:

voiced --> lenis

voiceless unaspirated --> fortis

voiceless aspirated --> fortis

English, French, Japanese:

voiced --> lenis

voiceless --> fortis

(French voiceless consonants except for tr, pr, cr clusters are usually unaspirated; Japanese non-initial syllables in a word are usually pronounced unaspirated for its voiceless consonants; English voiceless consonants are usually aspirated except after s- or o-, such as "stone" and "open", where it is then unaspirated)

[HashiriKata]

"Mandarin and Cantonese do not have voiced consonants."

The quote is out of context, and may mislead. The site is referring only to the "plosive consonants", not all consonants (n, m, r, z, etc are also consonants, and these are undoubtedly voiced).

The site's statement is correct in principle about the plosives but not quite, since those supposedly voiceless can become voiced in particular environments.

[HSC]

Ok, not that I am studying Shanghainese, but I would like to be able to hear the difference between the three categories, if someone knows how to make an audio recording and upload to this thread...

Voiced: 白，特，夹

Voiceless unaspirated: 百，德，革

Voiceless aspirated: 拍，脱，客

[carlo]

Another interesting thing is that native speakers of Mandarin usually can't hear the difference between voiced and unvoiced stops, so they will probably understand you even if you say /bu/ instead of /pu/ for 不, but if you listen closely you can spot the difference. I think there are regional variations too, Beijing speakers 浊化 more than speakers from other regions like Canton I think (but I'm not sure).

[ala]

The quote is out of context, and may mislead. The site is referring only to the "plosive consonants", not all consonants (n, m, r, z, etc are also consonants, and these are undoubtedly voiced).

Yes, sonorants (nasals, liquids, approximants) are naturally voiced; languages with voiceless sonorants are very rare. Mandarin Hanyu Pinyin z [ts] is not voiced though; r [z retroflex] is difficult to say as it also has sonorant character.

The site's statement is correct in principle about the plosives but not quite, since those supposedly voiceless can become voiced in particular environments.

Yes, for example the "de" in wode 我的 [uo d7] becomes voiced in Beijinghua. This voicing however does not occur in Cantonese and many southern, northwestern accents of Mandarin. In either case, this kind of voicing is not phonemic.

[ala]

Ok, not that I am studying Shanghainese, but I would like to be able to hear the difference between the three categories, if someone knows how to make an audio recording and upload to this thread...

Here's an audio sample (in Shanghainese, transcript below) distinguishing voiceless unaspirated, voiceless aspirated and voiced consonants.

Download Part 1 (800 kb)

ga go gei ji = voiced （茄、搞、勘、棋）

ka ko kei ci = voiceless unaspirated （加、高、减、机）

kha kho khei chi = voiceless aspirated （卡、靠、槛、气）

da do dei di （大、道、代、第）

ta to tei ti （带、岛、对、帝）

tha tho thei thi （泰、套、腿、舔）

ba bo bei bi （牌、袍、陪、啤）

pa po pei pi （巴、宝、杯、比）

pha pho phei phi （派、泡、配、骗）

tza tzo tzei tzi = voiceless unaspirated （债、早、再、资）

tsa tso tsei tsi = voiceless aspirated （差、超、菜、次）

nanenka / nanaka (怎样 how, what), khatsuo (卡车 truck), feiga (番茄：西红柿 tomato)

shipeinga (西班牙 Spain), pheifan (配方 formula), beilidochi (赔礼道歉 to apologize)

nieta (热带 tropics), thayan (太阳 sun), da (汏：洗 to wash)

ciseuci (计算机 computer), ciseuchi (计算器 calculator), jibeu (键盘 keyboard)

shiesitili (歇斯底里 hysteria), thielu (铁路 railway), dibi (地皮 land)

tzeiwei (再会：再见 goodbye), gontseitzini (共产主义 Communism)

More audio of Shanghainese voicing and sandhi (pitch accent) patterns:

Download Part 2 (400 kb) (transcript below)

sa su sa so shi （啥 苏 啥 少 死） = voiceless

za zu za zo zhi （惹 坐 惹 朝 徐） = voiced

ta tu ta to ti （带 多 带 岛 帝）

da du da do di （大 图 大 道 第）

sa su za zo zhi （啥 苏 惹 朝 徐）

za zu sa so shi （惹 坐 啥 少 死）

ta tu da do di （带 多 大 道 第）

da du ta to ti （大 图 带 岛 帝）

座位 zuwei (seat), 所谓 suwei (so called, such as)

葡萄 budo (grape), 群岛 junto (archipelago)

[Altair]

Compare the following English words: "pie," "spy," and "buy."

The "p" of "pie" is like pinyin "p," except that the latter has more aspiration (i.e., the consonant is released with a stronger puff of air).

The "p" of "spy" is the same as pinyin "b" and is different from the "p" of "pie." If you cannot notice the difference, hold your fingers right in front of your mouth and feel for the difference in the puff of air. The difference has to do with aspiration of the "p," which has little functional importance in English, but which is critical in Chinese.

The "b" of "buy" does not exist in standard Mandarin, but can be tolerated as a substitute for pinyin "b." Many Chinese that speak English use pinyin "p" for the sound in "pie" and pinyin "b" for the sound in "buy." This is part of what often gives such people a noticeable, but quite understandable accent.

From what I understand, the "voiced" "b" in Shanghainese is slightly different from the voiced "b" of English, which are both different from the voiced "b" of French.

The sound in Shanghainese comes from something called "murmur" that gives the entire syllable a breathy quality. It results from opening the vocal chords only partially.

The English sound comes from opening the vocal chords completely, but after the onset of the sound. In other words, voicing starts in the middle of the sound. This is true of the sound in initial positions, but not necessarily elsewhere.

In French, the sound is voiced from the very beginning.

By the way, as complex as all this sounds, the languages of the worlds have several additional varieties of these sounds. For instance, some languages combine closure of the lips with closure of the glottis to produce implosive or explosive airflow.

[ala]

The sound in Shanghainese comes from something called "murmur" that gives the entire syllable a breathy quality. It results from opening the vocal chords only partially.

This occurs only in the first syllable of a phonological word. All subsequent voiced consonants are truly voiced (nearly full opening of vocal cords) and have no murmur. The most obvious proof is the comparison of obstruent with sonorant behavior, for example obstruents 报道 [pO dO] and 报到 [pO tO]; both have the exact same tone sandhi pitch values, but 报道 [pO dO] is voiced on the second syllable. But in contrast sonorants 十一 [z@ jI?] and 十页 [z@ jI?] become homophones (!) in Shanghainese (while by themselves 一 [?jI?] and 页 [jI? Murmur+] are actually not homophonous), this is because only word-initial syllables can have the murmur; and since sonorants are naturally voiced and glottal stops disappear within a word, when in a non-initial position (hence without the murmur), the syllables will become homophonous. This shows that the distinction between 报道 and 报到 is not based on murmur nor tone, but rather on real voicing. While 十一 and 十页 are homophones because both 一 and 页 are sonorants which does not distinguish voiced or voiceless (one can say sonorants are always voiced) as soon as there is no murmur and when the pre-initial glottal stop vanishes.

In other words (sorry if above is confusing), what distinguished 一 from 页 was mainly the murmur; but there is no murmur in non-initial syllables, hence they become homophonous in 十一 and 十页. But 报道 and 报到 are not homophonous, because even if there is no murmur, there is still real voicing.

So Shanghainese is actually quite similar to the English, as it is also hard to actually characterize the murmur in Shanghainese (different conditions lead to different observations). Partial word-initial voicing and full voicing in non-initial syllables might actually be more appropriate of a description of Shanghainese voicing.

[HashiriKata]

Partial word-initial voicing and full voicing in non-initial syllables might actually be more appropriate of a description of Shanghainese voicing.

This is true, but since it is a phonetic (ie. physical) phenomenon, it's a universal feature with any language that has voiced plosives, isn't it?

The above feature doesn't therefore help much towards concluding that the Shanghainese /b/ is similar to the English /b/ (of course they are, relatively speaking). Since the plosives in Shanghainese belong to a 3-way contrast systems whereas those in English belong to 2-way systems*, they are very likely to have different phonetic qualities.

(* the voiceless unaspirated plosive doesn't exist in English as a meaningful member of a contrastive system, but only as an allophone of the aspirated member, due to the physical influence of the sibilant that precedes it.)

[ala]

The above feature doesn't therefore help much towards concluding that the Shanghainese /b/ is similar to the English /b/ (of course they are, relatively speaking). Since the plosives in Shanghainese belong to a 3-way contrast systems whereas those in English belong to 2-way systems*, they are very likely to have different phonetic qualities.

Yeah, it's hard to say that the Shanghainese voicing is identical to English. First, Shanghainese is still a tonal language, this complicates things a lot, and then what you said about its plosives being 3-way. Still, one can say with quite a bit of confidence that Shanghainese voiced consonants can and should be treated as separate voiced phonemes. So it's quite wrong to say that Shanghainese does not have voiced consonants but instead "voiceless consonants + murmur", because that is only valid for isolated monosyllable character pronunciations and clearly has a lot of trouble explaining multisyllabic behavior, where there is no perceived murmur in latter syllables. Even the exact notion of Shanghainese "murmur" is hard to pinpoint. Is it defined by the 1) onset? (as argued by Duanmu of Univ of Michigan); 2) the syllable?; 3) the nucleus? 4) the rhyme?; 5) the tone?; or 6) maybe the phonological word? (as argued by Zhu Xiaonong of the National Australian University). Several linguists and phonologists now believe it is indeed the phonological word that determines the Shanghainese murmur; making the murmur rather independent from voicing and other syllabic elements. That is quite interesting.

[HSC]

The audio files were interesting. Some words sound like a european language because the character pronunciations seem to run from one into the other producing a smooth and flowing word. This is unlike Mandarin in which you can usually hear the individual characters in a word, except for the 儿 suffix, of course

[Ferno]

yes HSC I see that too - to use a more Asian example you could use Japanese since it uses more European-like words rather than the Chinese style of mixing and matching different morphemes.. in fact some of those words sounded very Japanese!

on the subject of pronounciations slurring or running into each other...that's why I think European languages have a steeper learning curve at the begining in terms of listening

[hakkaboy]

Mandarin has the following voiced consonants: m, n, l, r, ng. All others are unvoiced.

[Desmond]

I noticed that nobody mentioned Hindi. It's quite interesting in that it has all 4 categories:

voiceless, unaspirated

voiceless, aspirated

voiced, unaspirated, and

voiced, aspirated!

That last one has always been amazingly tough for me, trying to make the "p" sound but putting voice behind it. Or alternately, making the "b" sound but trying to make it a plosive. I don't see how those Indians do it!

[ala]

Mandarin has the following voiced consonants: m, n, l, r, ng. All others are unvoiced.

These are sonorants (with arguable exception of r), sonorants are voiced in nearly all of the languages in the world. There are only a handful of tiny, obscure languages with voiceless sonorants.

When we talk about voiced consonants, we typically mean the plosives (stops) [b, d, g etc], fricatives [v, z, Z, z, h, G etc], and affricates [dz, dZ, dz etc]. Mandarin only has one then: r [z].

[Glenn]

I've been wondering about this for a few months: wouldn't it be more accurate to call the plosives/stops "plosives" when they're onsets, and "stops" when they're codas? It seems to me that they only explode when they're onsets, and they only just stop the air when they're codas. Does that make sense?

[Lugubert]

desmond:

...voiced, aspirated!

That last one has always been amazingly tough for me, trying to make the "p" sound but putting voice behind it. Or alternately, making the "b" sound but trying to make it a plosive. I don't see how those Indians do it!

You've gotta hear it. Think "energetic b". I suppose being close to Denmark makes it easier for me. The Danes have only voiceless consonants, and so it's the fortis/lenis difference for p/b.

In this context, I just have to point out how lucky we Swedes (and Norwegians) are, at least phonetically, when learning Putonghua or Hindi. We have no problem at all with the retroflexes (pinyin Xh, where X is c, s or z), or the plethora of Indian retroflexes.

[ala]

Swedes also have no problem with obscure vowels (like "oe" in Goethe and "o" in Swedish "bok") because Swedish has a ton of pure (monophthong) vowels!

[Lugubert]

ala:

Swedes also have no problem with obscure vowels (like "oe" in Goethe and "o" in Swedish "bok") because Swedish has a ton of pure (monophthong) vowels!

Yes, we're worse than most, including also the vowel of "7" (sju ) which not even Norwegians can pronounce correctly. For the IPA symbol of that one, no other language can be quoted than Norwegian and Swedish, but among us Scandinavians, we easily distinguish between the N and S varieties.

Back to Chinese, I obviously know that there is a n/ng problem, but is there anyone out there who can point me to a link or book or whatever giving an in-depth view on the pronunciations and geographical distribution of them?