Help talk:IPA for Mandarin

From Wikipedia, the free encyclopedia
Jump to: navigation, search


I know that [ɨ] is commonly transcribed [z̩] or [ʐ̩], but I have never heard these pronounced as fricatives, like English Zzzz. That is, Mandarin "four" sounds nothing like English says when said quickly enough for the vowel to drop out. Rather, they seem to just be very close vowels: We'd at least need a lowering diacritic in addition to the syllablicity diacritic. A transcription of [ɨ] has the advantage that ri is a rather intuitive [ʐɨ] rather than an obscure [ʐʐ̞̩]. kwami (talk) 08:54, 10 July 2009 (UTC)

Well, it really is [z̩] or [ʐ̩].

Historically, they were actually [i](about a thousand years ago), but because of some reasons, it was dropped and made consontants that before it syllable.

I don't know if I made it clear.Rethliopuks (talk) 17:36, 27 June 2011 (UTC)

English Approximation[edit]

The crucial problem with "English approximation" is that you don't specify what English/American/Australian/Canadian/etc. dialect you use. Senseless in the current way. —Preceding unsigned comment added by (talk) 22:32, 28 December 2009 (UTC)

Can you point out the examples where this is a problem? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:37, 28 December 2009 (UTC)
Possibly (?) it's referring to "song", which has ɑ for North Americans but not e.g. Britons. Lfh (talk) 14:45, 5 January 2010 (UTC)

There can be much improvement for the English approximation part. For example "X" sounds no way like "sh" but it is like the "xi" in taxi. There is a reason why pinyin chose "x" instead of "sh". (talk) —Preceding undated comment added 12:54, 25 October 2012 (UTC)

Not sure where you learned Pinyin, but that's false. The use of X for a sh sound probably parallels similar usage in Basque, Galician, Maltese, and Old Spanish. — Ƶ§œš¹ [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 14:34, 25 October 2012 (UTC)

Are the English Approximations something unique to Wikipedia, that is, something that has been provided by the general population, or is there a formal Chinese declaration on these approximations? I ask because, as a native English speaker, I fail to appreciate any difference between the CH in "cheer" versus that in "church" other than the following vowel sounds (ie; tɕʰ|q -vs- tʂʰ|ch). And since this is an apparently important distinction, I would think the approximations ought to reflect this. (talk) 09:45, 12 February 2014 (UTC)

There would not be a "formal Chinese declaration" on these approximations. Wisely, I think, government has stayed out of trying to tell, e.g., Basque speakers what sounds in their language are closest to sounds in Chinese.
Most people are not aware of the sounds that they make or how they make them. They are aware, instead, of the strange way another person makes "the same" sound in their speech. (I remember as a child frequently speaking with a lady who I now guess came straight from Ireland. Nobody even explained to me that she had an accent. I had to learn to listen to context and intuit what "correct sounds" she, in her strange tongue, was trying to make. I had, initially, no concept of how she produced some of those sounds, nor did it occur to me to try to understand how I made my "correct" sounds.) So the average speak of English is not the best one to describe what sounds in English are the best approximations for sounds in Chinese, and vice-versa. What you really want is a Chinese linguist such as Y.R. Chao (趙元任) who has really good English, or an English-speaking linguist like Jerry Norman who has really good Chinese.
Your own comment about "cheer" and "church" is a good example of our not hearing our own speech. I'm not sure where you come from. Maybe you have an old-timer's New Jersey accent, or some other regional variant wherein the two "ch" words are pronounced with the same initial consonant. When I position my tongue as though I'm going to say "cheer" and then try to say "church" I get a very strange sound. If, in school, I had pronounced "church" that way growing up in a town in Nebraska I would have been sent to treatment for my speech defect. I would be saying something that I can try to express in letters as "chee-urch." And if I tried to use the "ch" tongue preparation for "church" to say "cheer" it would turn into something coming from way back in my mouth and would sound more like "charch." That reminds me that I once knew a guy from Missouri who always talked about "jockey sharts," "boxer sharts," and so forth. He may have been making his "sh" sounds more like the Chinese retroflex "sh" sound. Too bad I didn't interrogate him at the time.
The use of English approximations is certainly not a "best practices" sort of thing. The "ch" of "cheer" is not the Chinese "q" sound. The tip of the tongue is in significantly different places for the two sounds, up in the airflow for English, and raising no turmoil for Chinese.


See [this site] for more discussion of what the tongue is actually doing for the "q" sound.P0M (talk) 16:43, 12 February 2014 (UTC)


Links to a sound's file should be for sounds that are consistent or that are close enough to the sound described in said article. Also, we shouldn't link to general articles on vowel length, tone, aspiration, or nasalization in the same way we do for phones. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 21:43, 22 February 2010 (UTC)

I see, Thank you.
See French page for nasalization, aspiration etc.
I will try to link the phones in the footnotes later on. (talk) 00:35, 23 February 2010 (UTC)

Minor quibble: Isn't the final <eng> rather pronounced [ɤŋ]? This fits with a symmetry between the allophones of other back vowels, such as [an]~[ɑŋ]. I'm not sure whether it sounds more like <sung> or German <jung> though.AlexanderKaras (talk) 03:55, 27 September 2010 (UTC)


I created a template to generate IPA from Pinyin. The conversion table could use a look-through—I probably butchered at least one of them. Also, any future changes to the transcription key here need to be modified there. See: Template:tone-cmn, Template:c-cmn, and Template:IPAc-cmn. Thanks! — ˈzɪzɨvə (talk) 05:15, 4 October 2010 (UTC)

many questions...[edit]

There are many questions I have here. It would be helpful if there were a list of sources for the values given in these tables. Quite a few of these, such as the "voiced" stops, reflect a very narrow transcription and should be sourced to someone's work. Rgr09 (talk) 12:10, 2 November 2010 (UTC)


Every so often someone thinks they can improve on Pinyin and tries to invent or impose a new system of transliteration. It's not a good idea. How many people are going to bother to learn this system in order to use Wikipedia? Pinyin is easily learnt and many people already know it. It's the standard means of transliteration of Mandarin into the Latin alphabet.

The system in use before was much better; the Chinese script and pinyin were easy to find at the start o the article. Now, if they're there at all, you have to hunt around in boxes and click links to find them.

If you don't want too much information at the start of the article, put the pinyin and characters there, and put the IPA in the infobox with the Wade-Giles and all the other obscure stuff which no-one actually uses in the real world. (talk) 09:28, 10 December 2010 (UTC)

It's a matter of accessibility. Not everyone knows pinyin, and pinyin doesn't mean the same thing with every language (do you know what ‹bb› or ‹mɡ› mean in pinyin?), whereas the IPA is universal, and should be familiar to anyone who can read a dictionary. Besides, we always use pinyin: it's the IPA that's secondary. — kwami (talk) 09:46, 10 December 2010 (UTC)

Minor changes[edit]

According to Mandarin phonology, the diphthong written [ou] here is actually [oʊ], and [ə̃] is [ɤ̃]. Which is correct?

I would also like a source on these voiced stops if possible. - AlexanderKaras (talk) 06:11, 12 December 2010 (UTC)

Whatever we decide, we should make sure that {{IPAc-cmn}} correlates with what we put here. That template seems to put tie-bars on affricates and semivowel diacritics for diphthongs when we don't (and shouldn't) have them here. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 06:27, 12 December 2010 (UTC)

call for deletion[edit]

Summary: I think that in it's current form, this article stands to do more harm than good, and probably doesn't have a place here anyway. I think it should just be deleted. Problems include:

1. The article poses itself as "the way in which the IPA represents Mandarin in Wikipedia articles," but that's not true, perhaps most importantly because Mandarin is not correctly represented by IPA in articles. It is correctly phonetically represented by pinyin. See Wikipedia:Chinese for more details.
2. I don't think a "sounds sort of like this English word but not really" list is particularly encyclopedic.
2b. Wikibooks has a considerably more useful and complete guide at:
3. Wikipedia already has an IPA equivalency chart at
4. The system claims to give English equivalents, but falls back to other languages frequently (unavoidably?) when no appropriate English sound is available.
5. There are no sources, and the whole thing looks like just a made up list to me. (This is probably related to point 2.)

I'm sure I could come up with more problems if I took more time, but my point here is not to dog on the article. My point, with all due respect to the authors, is that this article does not belong here and should be deleted. I'm posting here first rather than proposing a delete to see if anyone has any specific reasons it should be kept. — Preceding unsigned comment added by Carychan (talkcontribs) 15:05, 20 December 2010 (UTC)

I'll take your points one-by-one.
1. To say that pinyin is "correct" and Mandarin is "incorrect" is to apply mathematical values to what is really a stylistic difference of opinion. In your opinion, pinyin is better. However, one of the reasons we supplement pinyin with IPA is because it is not intuitive to our readers. We're already asking our readers to learn IPA for other languages, we shouldn't ask them to learn another transcription system if we can avoid it.
2. If you have a problem with the examples we've provided, there's nothing inherent about this page that says we can't use the ones you've offered. I don't know why bad examples would be "unencyclopedic." Maybe they're just bad.
3. That is an article, which is different. We need a place where readers and editors can both understand the IPA conventions for Mandarin to understand/implement them as well as to provide a central location (this talk page) to discuss issues of transcription that come about in one article that may affect others.
4. That is a problem, but, as I said for 2, is easily fixable.
5. Again, this isn't an article, it's a help page of sorts. This may seem like nit-picking, but it's an important distinction as it means that rules about sourcing are more lax.
While you've provided some valuable critique, you haven't demonstrated why this pronunciation guide should be deleted rather than fixed. Maybe looking at WP:DEL#REASON will give you some insight on the matter. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 15:31, 20 December 2010 (UTC)
I vote against deletion, a coherent IPA system is important in learning how to pronounce Pinyin correctly. Any article that promotes information about how pinyin should be pronounced is a good idea. Pinyin is not phonemic, it's just a romanisation. And it's really difficult to know how to pronounce pinyin. More Mandarin textbooks, dictionaries and language resources should use the standard phonemic system (IPA) to describe pronunciation. When you are studying any other language, which is written in Latin characters (such as English, Spanish, French or German) you have the IPA pronunciation next to the word written in Latin characters. Also when you are studying Mandarin, you should have the IPA pronunciation next to the pinyin word written in Latin characters.
2b. Wikibooks has a considerably more useful and complete guide at:
To my mind, that doesn't seem very useful, except for English speakers...
For the reference, I'm not native English speaker, though I've studied it and several other languages quite a lot. I'm a HSK certified level 2 Mandarin speaker so I have some experience, though there's a long way still to go. And I really appreciate all IPA resources to help how to understand and pronounce pinyin. Panu-Kristian Poiksalo (talk) 15:34, 1 May 2011 (UTC)

Remove bopomofo?[edit]

Recently, bopomofo was added, but it does not represent individual phonemes but entire finals, hence it doesn't seem appropriate here. Should be removed in my opinion. Officer781 (talk) 09:01, 7 June 2011 (UTC)

Does that mean that the table at bopomofo is also wrong? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 11:40, 7 June 2011 (UTC)
It has finals where the pinyin has finals. — kwami (talk) 02:05, 8 June 2011 (UTC)


Hi, I was directed here from Chang'e 1, which gives the pronunciation as [tʂɑ̌ŋ.ɤ̌], but the character ɤ is not listed here, only ɤʊ. Not sure which needs correcting... (talk) 10:43, 28 August 2011 (UTC)

Sorry, an editor's been changing the templates, but inconsistently so that the IPA is not properly supported. — kwami (talk) 06:33, 19 September 2011 (UTC)

section "p b span"[edit]

p b span

I have concerns about this line. I think this concern is about the majority of sounds marked as "b" in pinyin. An example is the number 8. In Chinese this is "bā", and is pronounced [pʷa˥] or simply [ba˥] when not fully voiceless by context. The English gloss suggests that to English speakers this is "pa", like the p in "span." But it is not at all, it is very labialized, and rounded, and sounds far more like [b] as in "ban". It is so close to a [b], I would guess that nine out of ten native English speakers would fail to recognize that it is a [p] at all. I suggest a solution like adding a new line:

b ban

Or, change the line to read

pʷ p b, b ban, span
The difference between the p of span and the b of ban is not one of labialization. We would only want to include a line with [pʷ] if we transcribe words with that. Do we? — Ƶ§œš¹ [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 16:18, 24 March 2013 (UTC)

zhi, chi, shi, zi, ci, si, ri[edit]

The article Close_back_unrounded_vowel says it's ɯ instead of ɨ.-- (talk) 19:13, 2 September 2013 (UTC)


I'm a native speaker grown up with German. When I pronounce sounds like pinyin sh/ch, they just sound like the English sh/ch or German sch/tsch. I never knew there was a difference before I got to know IPA. If I really pronounce them retroflexly, it sounds exaggerated as if I had a strong Northern dialect which isn't standard Mandarin anymore.-- (talk) 21:44, 4 September 2013 (UTC)

Mandarin retroflexes are laminal, so they're not retroflex in the 'true' sense. The difference ought to be that they're not palatalised, whereas German and English sh/ch/etc. are somewhat palatalised. — Lfdder (talk) 22:34, 4 September 2013 (UTC)
Actually, German postalveolars often sound just like the Polish ones, i.e. are laminal retroflex. Peter238 (talk) 11:51, 29 July 2015 (UTC)


Is [v] standard? Because even news anchors do that.-- (talk) 23:07, 19 September 2013 (UTC)

No. It's a recent "import" from other regional Chinese "dialects." It may become permanent, or it may disappear when it goes out of vogue. (Do you remember when a radio announcer from GB introduced a form of English pronunciation wherein "s" such as the initial consonant in "sob" were all turned into "sh" sounds? I think the practice even extended to sentences such as, "I usually sit on the davenport." Thank goodness that fad self-extinguished. Who knows what will happen to "v" in Mandarin. I once had a philosophy teacher in Taiwan who was from somewhere in Zhejiang. He spoke of "物理學 vǔ lì xioh*" (with xioh in 入聲). It didn't bother my Chinese schoolmaets. They acted as though they didn't notice his pronunciation even when he referred to "thinkers" as 死像家, which seemed riotously funny to me. Nobody else even cracked a smile. The thing about Northern Chinese is that lots of people like my old teacher moved around, especially (I suspect) in the Great Cultural Revolution. Maybe the ones who ended up in the north had high status? For some reason their "new speech" attracted many converts for "v." P0M (talk) 16:56, 12 February 2014 (UTC)

Nasal vowels[edit]

I believe there are no real nasal vowel in Mandarin Chinese. "ang", "eng" and "ong" should be [ɑŋ], [əŋ] and [ʊŋ].-- (talk) 19:29, 30 October 2013 (UTC)

Questionable values[edit]

The word 安 does not sound like the English word "an", so the vowel should not be represented by [a]. [ɐ] would be better.

English "an" isn't represented by [an] but [æn].-- (talk) 18:26, 1 December 2013 (UTC)
The difference between [a] and [æ] is that [a] is "open" and "[æ]" is halfway to "open-mid", so the difference between the two is very slight. [ɐ] is "open-mid" and central, so it isn't quite like [a] either. A small region in the U.S. uses [a], and the discussion at Open_front_unrounded_vowel along with the pronunciation sample at

give something very different from the Chinese sound.

ʊ in, e.g., 東 tʊŋ is explained as sounding like the "oo" in "hook," but "oo" is a little too rounded. Maybe there is no better choice.

I don't see the [ɯʌ] (or just [ʌ]) needed to give the pronunciation for 德.

Where did you read that? It's [ɤ].-- (talk) 18:26, 1 December 2013 (UTC)
What makes you an expert? Seriously, I am not asking a rhetorical question.
Check and use "voice" IPA2.
Check Close-mid_back_unrounded_vowel and hit the play button in the taxo box there.
I can't find any English words that are transcribed with [ɤ].

An example is given for [ʌ] in footnote 1o for "some value [ˌsʌɱˈvæɫjuː])" in International_Phonetic_Alphabet_chart_for_English_dialects, and it certainly sounds better than your choice.

I don't see the [ɪ] needed to give the pronunciation for 民 [mɪn]. (It is not pronounced like the English word "mean" unless someone is hyper-correcting and trying to work out the sound s/he should be producing to correspond to the dictionary's "min".

Again, English "mean" isn't [mcn] but [iː].-- (talk) 18:26, 1 December 2013 (UTC)
You misunderstand. The Chinese 民 is being transcribed as [mi:n], but Chinese people don't pronounce it that way unless they are over-correcting because of pinyin transcriptions. 一 yi but 民 min. The two "i" sounds are different. [ɪ] is the sound in 民 and in English kit, lid, fill, bin.
The article on pinyin that provides IPA equivalents for various pinyin "spellings" offers
yin        bin     pin     min     nin     lin     jin     qin     xin

Those IPA values ought to produce pronunciations that would sound like the vowels in English:

Een        bean    peen    mean    neam [oil]      lean    jean  cheap sheen
              ball and peen 

The vocal tract needs to be more relaxed, as with [ɪ]

(additions above to bring discussion forward) 01:27, 13 February 2014 (UTC)


From what I've heard from Mandarin pronunciation words syllables that start with "h" (Pinyin) don't seem all that velar. At the very least it's an /x/ that is audibly different from, say, German or Russian. And certainly very different from "loch" (as given in the table).

At the very least, it seems to be pronounced quite differently in "han" than in "hu" or "he". Does anyone have a detailed, reliable phonology source regarding /x/ in Mandarin?

Peter Isotalo 00:03, 9 December 2014 (UTC)

I'll be cleaning up that page. It's not like "loch" because "loch" has is much harsher like [χ] instead of [x]. But in broad transcription, it doesn't make any difference. The standard pronunciation for pinyin "h" is [x]. This is especially prevalent in the North where the newsanchor pronunciation comes from. In the South it might be less harsh and reduced to [h] because of the influence of Wu and Cantonese. As you've mentioned it can also depend on the vowel following pinyin "h". But [x] is generally correct. -- (talk) 18:21, 21 December 2014 (UTC)

English Approximation for tʂ, tʂʰ, ʂ, ʐ[edit]

Is there any reference or citation? I am Native Mandarin speaker, and for these four sounds, I don't really think there is any English Approximation for tʂ, tʂʰ and ʂ.

By far, there is only Russian Approximations: Ш for ʂ and Ж for ʐ; in addition, ʐ sometimes does has English Approximation "Raw" when combined with certain vowels.

As for examples,

  • tʂ is not roughly like "jaw" [ɔ], which is similar to edgyɪ], more like tɕ (without -r sound), which is totally different from tʂ (with -r sound).

Like wise,

  • tʂʰ is not roughly like "church" [ɝtʃ], which is similar to cheat [[it], more like tɕʰ (without -r sound), which is totally different from tʂʰ (with -r sound).
  • ʂ is not roughly like "show" [ʃo], which is similar to she [[ʃi], more like ɕ (without -r sound), which is totally different from ʂ (with -r sound).

They are not "roughly like" but totally different.

I suggest we should mark tʂ, tʂʰ, ʂ with [No English equivalent] like some entries in Help:IPA for French.

--TX55TALK 16:09, 2 May 2015 (UTC)

Possible mistakes[edit]

In the table it reads

tɕ j roughly like edgy
tɕʰ q roughly like cheat
ɕ x roughly like she.

Shouldn't these be

tɕ j roughly like edgy
tɕʰ q roughly like cheat
ɕ x roughly like she?

K9re11 (talk) 13:14, 20 May 2015 (UTC)

Removed three vowels not produced by the template[edit]

Also reduced the entries to sequences that produce the phones in question. — kwami (talk) 22:15, 28 July 2015 (UTC)

I understand I made and removed a complaint yesterday about the over-precision of a. However, it seems this is standard practice in chinese phonology writings and my edits were reverted at the standard chinese phonology article as well. Hence I have reverted the removals.--Officer781 (talk) 02:53, 29 July 2015 (UTC)
I'm for removing [ä] from this guide, but not from Standard Chinese phonology. We don't have to list every single allophone here. That seems to be the general policy for the IPA guides (but not phonologies), and that's how it is on Help:IPA for Russian, where the open vowel /a/ is transcribed [a], rather than [ä], the mid vowels are written [e, o] instead of [e̞, o̞] (even though both close-mid [e] and mid [e̞] occur as allophones of /e/), and many allophones ([i̝, ɪ̝, ɨ̞, ʉ̞, e̠, e̽]) are simply not listed. The same kind of simplification applies to Russian consonants. Peter238 (talk) 11:40, 29 July 2015 (UTC)

Removed them again. They're not generated by the template, so it is incorrect to include them here. — kwami (talk) 04:37, 30 July 2015 (UTC)

Changed the generic a to a central vowel. I'm not sure how the template works for "generated" but I'd suggest sticking to an external source on transcribing the sounds (in this case using the generic IPA symbol for central "a" as the underlying phoneme as how the standard mandarin phonology books and papers out there do it).--Officer781 (talk) 14:11, 30 July 2015 (UTC)
I wonder about the recent changes. I've never heard ei or ou pronounced with a mid-central vowel, at least not that I've noticed. — kwami (talk) 04:24, 31 July 2015 (UTC)
It's a rough transcription that's in the literature that uses the underlying vowel to transcribe (otherwise we would have too many sounds to use here with four variants of a). For example, English [ʌ] is almost never pronounced this way anymore by nearly every variant including GA and RP (pronounced as [ɐ] instead), but is retained for various purposes. I feel WP's IPA pages should be used more as a rough guide on how to pronounce rather than an "exact" pronunciation which would be covered in the phonology page instead. Since Wikipedia seems to mainly use Duanmu (2007), it's in the appendix. See the link I gave you on your talk page.--Officer781 (talk) 13:35, 1 August 2015 (UTC)

"May be pronounced with a light frication"[edit]

I just noticed that Officer781 reverted ‹See Tfd›themself regarding the note "May be pronounced with a light frication". This does not seem useful for the purposes of this page: This page is a HELP page designed to help readers pronounce Mandarin, as a simpler alternative to Standard Chinese phonology. What benefit could there possibly be in discussing phonological variations that none of the intended audience needs to worry about? — Sebastian 19:26, 19 December 2015 (UTC)

Ok, removed the note.--Officer781 (talk) 02:25, 20 December 2015 (UTC)

Move discussion in progress[edit]

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:17, 15 July 2017 (UTC)