A Little Primer of Xinjiang Proper Nouns

« previous post | next post »

Following the serious unrest in the Xinjiang Uyghur Autonomous Region of the Peoples Republic of China that erupted last week, "Xinjiang," "Uyghur," "Urumchi," and other names pertaining to the region have become household words.  Unfortunately, people around the world have been confused about how to pronounce these words.  This is understandable for several reasons.  In the first place, we have to confront both the original Uyghur terms and their full and truncated versions in Mandarin.  Second, there is often a plethora of variant romanizations for each name.

Ed Wong, the New York Times correspondent who has been keeping us so well informed about the events as they unfold, told me that the NYT had

"received an email from a reader saying the NYT should change its 'pronouncer' on Uighurs.  Right now, in our articles, the editors insert (WEE-gurs) as the pronouncer.  One reader said this is not the correct pronunciation, and sounds strange to the Turkic speaker’s ear."

As I told Ed, it would have been helpful if the person who sent the NYT the e-mail would have indicated the correct pronunciation, not just told them that WEE-gurs sounds wrong.

In an attempt to clarify how the most important  Xinjiang names are actually pronounced in Uyghur and in Mandarin, I here provide various orthographic forms along with audio clips.  As to what sort of "pronouncers" should be developed for the major media, presumably to represent appropriate Americanized pronunciations, I invite suggestions.

[Update: I give "pronouncers" and recordings of suggested American English renditions in a later post, "American English pronunciation of Uyghur proper nouns", 7/15/2009.)

The names are read first in Uyghur and then followed by their Mandarin versions.

1. Uyghur — Wei2wu2'er3     維吾爾 /维吾尔      ئۇيغۇر
Variant romanizations of the Turkic include Uighur, Uigur, and Uygur.  I prefer Uyghur as being closest to the Perso-Arabic orthography.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

2. Ürümchi — Wu1lu3mu4qi2 烏魯木齊/乌鲁木齐   ئۈرۈمچی
Variant romanizations of this Uyghur word for the capital of the region include Urumchi, Urumtsi, and Ürümqi.  The latter is the official cartographic spelling in China; it's a combination of Turkic and pinyin.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

3. Täklimakan — Ta3ke4la1ma3gan1 塔克拉瑪干/塔克拉玛干   تەكلىماكان
Variants:  Taklimakan, Taklamakan. This is the huge desert that occupies most of southern Xinjiang; it is one of the largest and most arid deserts in the world.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

4. Tarim — Ta3li3mu4 塔里木   تارىم
The name of the basin in which the Täklimakan Desert sits.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

5. Tarim River (Daryasi) — Ta3li3mu4 He2 塔里木河   تارىم دەرياسى
The long river flowing along the northern edge of the Täklimakan Desert from which the Tarim Basin derives its name.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

6. Kashgar — Ka1shi2ka1'er3  喀什喀爾/喀什喀尔   قەشقەر
Variant:  Qäshqär.  The large oasis city at the far western edge of the region that is famous for its Sunday bazaar, the largest mosque in China, and other important monuments.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

7. A shortened Mandarin version of no. 6.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

8. Tängri Tagh — Tian1 Shan1 天山   تەڭرىتاغ
Variant:  Tien Shan.  The Celestial / Heavenly Mountains that divide the northern part of the region from the south.

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

While it cannot be expected that this post will instantaneously clear up all the confusion surrounding the pronunciation of proper nouns related to the situation in Eastern Central Asia (recall how hard it was just to deal with Beijing last year during the Olympics:  "How they say 'Beijing' in Beijing"), I hope that it might serve as a reference for those who strive for accuracy.

Incidentally, Xinjiang (Xin1jiang1 新疆) means "New Borders," "New Territories," or "New Frontier," and is the name that was given to the region after it was subdued by the Manchus in 1884 and incorporated into the empire of the Qing Dynasty.  To avoid using politically sensitive names such as East Turkestan or Uyghurstan, I refer to the region by the purely geographical designation as "Eastern Central Asia."

My thanks to Dolkun Kamberi for providing the Uyghur and Mandarin recordings.


  1. language hat said,

    July 13, 2009 @ 7:31 pm

    One reader said this is not the correct pronunciation, and sounds strange to the Turkic speaker’s ear.

    I will issue my standard disclaimer that English spellings and pronunciations are for the use and convenience of English speakers, and it is foolish and presumptuous to expect them to sound correct to speakers of other languages. I seriously doubt that a Uyghur speaker's rendition of, say, "New York" would pass muster to an English speaker, and that's as it should be. Different languages are different.

  2. Nathan said,

    July 13, 2009 @ 7:40 pm

    What, no IPA transcriptions?

  3. dw said,

    July 13, 2009 @ 8:04 pm

    I can't listen to the audio examples in my browser. Would someone mind transcribing them in IPA? Thanks.

  4. Bloix said,

    July 13, 2009 @ 8:36 pm

    If the Turkic pronunciation really is "weewer" then why the heck is it transliterated as Uygher? "Uy" is bad enough – initial "U" can be "Yoo" or "Uh" but it's never "W" in English – but at least we can make some sense of "U" as "Ooo" and "y" as "ee" – so "oo-ee" beomes "wee." Pretty bad, but not hopeless. But "gher" for wer? What is that? Is this a game that only Perso-Arabic specialists are allowed to play? What on earth would be so bad about Weewer?

    Seriously, the quickest way to lead people to believe that foreign events are unintelligible is to make it impossible to keep the names of people and things in mind. And monolingual people simply can't retain groups of letters that can't be sounded out and recognized.

    If you read a news story that began, "Today in Lsydkf the Ksdfmo people rioted against the Jmaksfxid goverment in their capital city, Qotnsd," how much farther would you read? Well, maybe as a linguist you'd be fascinated, but ordinary people will just give up. They really don't care which othography comes closest to the Perso-Arabic. They want to be able to read the paper, and the spelling Uygher gets in the way.

  5. John Biesnecker said,

    July 13, 2009 @ 9:27 pm

    Fantastic post, Victor.

    Of all the proper nouns revolving around the Xinjiang protests, I think the one that gets butchered the worst is "Xinjiang" itself. I've heard everything from almost correct pronunciations to newscasters (I want to say it was one of the BBCMundo Spanish broadcasters) going all Wade Giles and delivering "Sin Kee-ang." It's been painful.

    I'm curious, though, how do you say "Xinjiang" in Uyghur? As you say, its a politically loaded word, but there must be a representation, at very least in Uyghur-language Chinese propaganda.

  6. Mark A. Mandel said,

    July 13, 2009 @ 10:15 pm

    The pronunciation links don't come through on LJ. I put a comment there to that effect, adding "Click on the source link to see and then listen to them."

    Bloix: Perhaps this is the first time you've ever seen the name "Uyghur", but it's not new at all. I don't know where the spelling came from — maybe Victor can tell us — but that's what we've got, and we're pretty well stuck with it, like everything else in our orthography. Most everything that's ever been written in English about this language has probably used this spelling or one like it.

    Hey, what's with that "ai" in "Britain"? Nobody says "brit-AYN". From now on I'll write "Britn". And "Rusha" and "Beljum". And "Noo Yawk" and "Kaintuck" and "Bawlmer"……..

  7. Garrett Wollman said,

    July 13, 2009 @ 10:30 pm

    @ Mark, Bloix:

    That "gh" in the Uyghur pronunciation of "Uyghur" sounds suspiciously like the sound that "gh" used to represent in English….

  8. Tom said,

    July 13, 2009 @ 10:40 pm

    Uyghur / ئۇيغۇر – [ujɣur]

    Ürümchi / ئۈرۈمچی – [yrymtʃi]

    Täklimakan / تەكلىماكان – [tæklɨmakan]

    Tarim / تارىم – [tarɨm]

    Tarim Daryasi / تارىم دەرياسى – [tarɨm daryasi]

    Kashgar / كاشگار -[kaʃgar] or Qäshqär / قەشقەر – [qæʃqær]

    Tängri Tagh / تەڭرىتاغ – [tæŋri taɣ]

    (If someone knows better than I do, feel free to correct these; just a rough and dirty transcription.)

    A lot of these spelling and romanization problems have to do with the politics of writing. Both the Chinese and Russian governments have introduced at least four writing systems over the years, based off of the Arabic, Cyrillic, and Latin alphabets. These are on top of the Chagatai script, derived from the Perso-Arabic script, the Uyghurs had been using themselves.

  9. Tom said,

    July 13, 2009 @ 10:46 pm

    @ Mark:

    It is "Bawlmer" in Bawlmerese, hon.

  10. Bloix said,

    July 13, 2009 @ 11:39 pm

    Actually, because I'm a lawyer interested in civil rights in general and Guantanamo in particular, I've known about the Uighers for several years now. The spelling I'm used to is Uigher. I find Uygher annoying but tolerable. But what I find ridiculous is that I'm supposed to say "weewer." I do understand that we need to have a transliteration that is reasonably uniform across European languages. But Uygher for weewer? What language does that work in?

  11. Jon Lennox said,

    July 13, 2009 @ 11:49 pm

    But where does Weezer come into it?

  12. Lazar said,

    July 14, 2009 @ 12:05 am

    @Bolx: There are many different spellings used in English – Uyghur, Uighur, Uygur, Uigur – but AFAIK you need to use a "u" in the second syllable, not an "e". And it's not "weewer" that we're supposed to say, but "weegur".

  13. Bob Violence said,

    July 14, 2009 @ 12:06 am

    "Uyghur" is the name of the people (and the language) as written in the "Uyghur Latin Yéziqi", developed in the early 2000s to serve as the new standard Latin script for the Uyghur language. The "gh" digraph, as Garrett correctly assumed, corresponds to ɣ — note that this also true of the "gh" in "Afghan". In the former Latin system (used until 1987) it would've been "Uyƣur." This is also where the still-official "Ürümqi" spelling comes from, as the old system used "q" for ʧ(by rough analogy with Hanyu Pinyin) — in the new system it's "ch".

  14. Etl World News | UIGHUR NAMES PRONOUNCED. said,

    July 14, 2009 @ 12:07 am

    […] Mair has provided an invaluable post over at the Log, giving audio clips of "how the most important Xinjiang names are actually […]

  15. Bob Violence said,

    July 14, 2009 @ 12:18 am

    Addendum: Here's a neat little paper on the history of Latin scripts for Uyghur and the development of the new script. For John, this paper gives the Latinized Uyghur for "Xinjiang" as "Shinjang", which might correspond to [ʃinʒaŋ], although I know nothing about Uyghur phonology so I can't say for certain.

  16. Steve Politzer-Ahles said,

    July 14, 2009 @ 12:56 am

    @ Tom: your IPA transcriptions are accurate, except that [tarɨm daryasi] should be [tarɨm darjasi] (honest mistake!). The "i"s at the end of words, such as in Urumqi, are often more of a central /ɨ/ than a real /i/, but that's just a small quibble (and probably varies by speaker; I'm learning from an Uzbek Uyghur so her pronunciation might be a bit different than someone from Urumqi). And the "gh" is often a bit like a trill /ʁ/, although not so flappy; so both [ujɣur] and [ujʁur].

    As for an easy pronunciation guide for English speakers… the way I tell my parents and friends to pronounce it is "oi" (like a stereotypical Jewish person?) + "gore" or "oi" + "whore".

  17. Steve Politzer-Ahles said,

    July 14, 2009 @ 12:59 am

    Oh, also… for Kashgar, which you transcribed as

    Kashgar / كاشگار -[kaʃgar] or Qäshqär / قەشقەر – [qæʃqær]

    I have never seen the first (either that pronunciation, or that spelling in the Perso-Arabic script), only the second, with the uvulars and all that fun stuff. I think Kashgar is just the way we spell it in English?

  18. Bloix said,

    July 14, 2009 @ 1:35 am

    Lazar, listen to the first clip. There's no "g" sound in either pronunciation. You've got Wee-wer (Uyghur) and Wee-War (Mandarin). And okay, it's a "u." Why? The usual ending in English is "er." If you've got the h, there's no chance of a soft g, so why make it a "u" other than to be difficult?

    And Steve P-A, NO ONE but NO ONE says Oy-hoor. You are making stuff up.

  19. Steve Politzer-Ahles said,

    July 14, 2009 @ 1:44 am

    @Bloix: I wasn't suggesting that that's how people around here do pronounce it (thanks to the popularization of the "weeger" pronunciation aid). I was just suggesting that it's a relatively close approximation of the native pronunciation, using sounds that exist in English. It's really not worth fighting over, though, so I'll leave it at that.

  20. Gao said,

    July 14, 2009 @ 4:54 am

    Tarim Däryasi (not “Daryasi”) / تارىم دەرياسى – [tarɨm dæryasɨ]

    Qäshqär / قەشقەر – [qæʃqær]: the official Romanized name is “Kaxgar”. كاشگار is Farsi, قەشقەر is Uyghur.

  21. jo said,

    July 14, 2009 @ 7:34 am

    @Bloix: But Uygher for weewer? What language does that work in?

    This complaint seems to me to be wrong both in substance and in intent.

    Firstly, it seems you are having trouble hearing the sounds of the word Uyghur, but if we accept that the phonetic transcription [ujɣur] is basically correct, then you don't have to understand IPA (or to have a good ear for the sounds as you hear them) to see that the letters in the word Uyghur actually correspond fairly well to the sounds of the word as transcribed.

    Secondly, I think you are making an essentially unfair criticism of the Uyghur romanisation system(s). You accuse the romanisation of being intentionally 'difficult' and phonetically opaque for English speakers. But the goal of making a romanisation that is easy to read and pronounce for English speakers is just one of many competing demands that a romanisation system has to balance. Others might include matching roman letters to letters in the original script, being phonetically and/or phonemically consistent, or accommodating the needs of speakers of langauges other than English.

    In addition to the above, I think languagehat's point at the top is very right.

  22. J. W. Brewer said,

    July 14, 2009 @ 7:38 am

    I'm assuming the newly-codified preference for Uyghur over Uighur reflects the same mindset that decided the former Kirghiz SSR ought to be spelled Kyrghyzstan? Bad PR judgment, unless your goal is to discourage Anglophones from paying any attention to you and your regime's record on human rights. Whatever else you may say about those who seek independence for "East Turkistan," they know how to appeal to Anglophones.

    It's a pity the old toponym Dzungaria seems to have fallen into desuetude. That was a nice one for Anglophones: suitably exotic, but not impenetrable.

  23. J. W. Brewer said,

    July 14, 2009 @ 8:06 am

    In light of Jo's comment, I should perhaps make an implicit distinction explicit. It's presumably a good thing for there to be standardized Latin-alphabet spelling conventions for Uighur that can be used by those wishing to read or write that language in that script and which can at least be a starting point for those needing to quote Uighur words in texts in other languages standardly written in the Latin alphabet. As Jo correctly notes, being transparent or intuitive for English-speakers is not the sole or dominant criterion for such a set of conventions. But those conventions should not of their own force govern what the English toponym, ethnonym, etc. related to that speech community ought to be or how those English words ought to be spelled. It's the difference between Srbija and Serbia.

  24. Mr Fnortner said,

    July 14, 2009 @ 8:45 am

    Bloix's lament has received unjustified ridicule here, and you linguists need to get down from your high horses for a few minutes and look at yourselves and what you are perpetrating.

    Perhaps the lector who pronounced Uyghur for our benefit did not enunciate well enough, but her first word, the one intended to be spoken in Uyghur, is definitely pronounced 'weewer." If you can hear a "g" of any sort between wee and wer, then you are hallucinating it. Before you castigate Bliox any further on this point, provide a recording from which he and I can hear the g.

    Second, many people (apologies, Bliox, no offense intended) do not have a facility for synthesizing diphthongs from unexpected vowel combinations in written words or, in reverse, recognizing that the letter w is almost always used to represent the sound of u, or that y is pronounced as the letter i, regardless of how we were taught to read w and y. So it is no wonder that Bliox does not see "wee" in u-y. This misapprehension was an opportunity to reveal this to him, not to dismiss him as some sort of peasant.

    Third, his point about alien transliteration systems is well made, and professional linguists could do the world a service through a concerted effort to improve the quality of representation of the sounds of one people's language in the alphabet of another's. Not every system is as refined as the romanization of Japanese, for example. The current system used for Chinese thwarts every effort for even college educated, moderately diligent English speakers to pronounce their words with any accuracy. It is as if the Chinese intended we would not be successful when they chose that system.

    Basically, the dialog has gone like this–
    Bliox: "I hurt."
    Linguists: "Of course you do. You're a cripple. Go back to your room."

    That was shameful. It would have been better had you diagnosed Bliox's pain, then taken your substantial skills and helped him overcome what was bothering him.

  25. Nightstallion said,

    July 14, 2009 @ 8:57 am

    "If you can hear a "g" of any sort between wee and wer, then you are hallucinating it."

    Nonsense – it's plain as the sky that the first pronunciation is [uigu:r], as in Buick (without the [y]) and game and rural, so I see no reason why you should protest that the letters and sounds are too strange and non-understandable.

  26. Franz Bebop said,

    July 14, 2009 @ 9:04 am

    The "pronouncer" is actually called a "respelling," right?

    Letters don't pronounce. People pronounce. :-)

  27. Jon Lennox said,

    July 14, 2009 @ 9:26 am

    "If you can hear a "g" of any sort between wee and wer, then you are hallucinating it."

    In the Turkic version, I don't hear a "g", I hear a "ɣ", which I think is the sound you're transcribing as the second "w".

    The "ɣ" sound was traditionally written "gh" in English, back when it had the sound, much to the detriment of subsequent spelling ("enough", "though", "through", "eight") once the sound dropped out of the language.

    Despite the English spelling, pronouncing "ɣ" as "g" is pretty inaccurate, though. Anyone have suggestions for a more appropriate closest English phoneme?

  28. J. W. Brewer said,

    July 14, 2009 @ 10:06 am

    Back to square one. VM invited suggestions as to appropriate "pronouncers" for English-speakers. He implicitly assumed that information as to how the words are pronounced in Uighur and Mandarin would be useful toward that end. It does seem that data ought to be relevant, but if it turns out to be the case that the echt-Uighur pronunciation involves a consonant that does not currently exist in English, we seem to be stuck in terms of recommending the optimal English pronunciation, unless it turns out to be the case that there's a consensus on which existing-in-English consonant is the closest substitute. Failing such consensus, using the Engliish hard "g" seems the most straightforward thing to do since it's the closest fit to the range of variant English spellings, and it's not like an English "w" (or a raspy sort of "h"?) would lead to the authentic but impossible-in-English pronunciation if what's really needed is the IPA-gamma.

    If only our ancestors hadn't given up the yogh! (A misguided attempt to make things easier for foreigners?)

  29. Franz Bebop said,

    July 14, 2009 @ 10:15 am

    @jo: But the goal of making a romanisation that is easy to read and pronounce for English speakers is just one of many competing demands that a romanisation system has to balance. Others might include matching roman letters to letters in the original script, being phonetically and/or phonemically consistent, or accommodating the needs of speakers of langauges other than English.

    This little debate reminds me of the debate over spelling Akhilles vs. Achilles. The spelling Akhilles always bugged me. A "k" is not a kappa, a "k" is a "k", so there's nothing implicitly more authentic about "k" rather than "c" for representing kappa. Same thing with preferring "kh" over "ch" for the letter chi. What difference does it make? If you want authenticity, then just use the letter chi directly.

    Jo is correct, of course, but what Bloix seems to be advocating is not Romanization but Anglicization. Why should the spelling of the English name for "Uyghur" match the French spelling? I don't see a compelling reason for this. What good is a consistent Romanized spelling if huge blocks of people can't make use of it?

    Californians can name their state whatever they want, but they cannot command the Germans to stop referring to the state as Kalifornien. For the same reason, the Serbs should not try to tell English speakers to write Srbija rather than Serbia. Srbija may be authentically Serbian, but it's not authentically English.

    What's the Romanization for Россия ? Is it Russia, Russland, Russie, Rusko, or Rusya? Should we insist that the Germans write "Rossija" instead?

    If IPA is important, then just use IPA directly. "Uyghur" is not IPA.

    In English, "Weewer" is as good a spelling as any. "Oyhoor" might even be better. It doesn't matter to me how speakers of other languages spell the name. It would be wise for every language community to choose a spelling that tends to elicit from its speakers a pronunciation as close as possible to the original. It's also important to choose a spelling that is meaningful and usable. But universally-consistent Romanization does not seem to be a useful goal.

  30. Saqer A (saqeram) 's status on Tuesday, 14-Jul-09 14:15:56 UTC - Identi.ca said,

    July 14, 2009 @ 10:17 am

    […] the linguistic perfectionists on pronouncing #Xinjiang names like #Urumqi: http://languagelog.ldc.upenn.edu/nll/?p=1576 #language […]

  31. vanya said,

    July 14, 2009 @ 11:02 am

    Sorry – "Weewer" is ridiculous. Firstly, there is no "w" in the second syllable, it's a very clear "ɣ" sound. Secondly, Uyghur is a long established spelling (1875 according to a Google book search, maybe older). Thirdly "weewer" is to close to the Mandarin pronunciation and would doubtless be found offensive by Turkic speakers in general, not to mention the fact that it just looks a child's potty word. Finally, I would guess that 98% of the people who actually have any interest in Eastern Turkestan do just fine with Uyghur/Uighur. As I ask every time enthusiasts want to "simplify" spellings, change alphabets, rename places, etc. – what will this reform actually accomplish? cui bono?

  32. Franz Bebop said,

    July 14, 2009 @ 11:23 am

    If "Uyghur" is long-established, I would accept it as a strong argument for keeping it as-is.

    Ridiculous is in the eye of the beholder. The reason this mini-debate got started is because "Uyghur" also seemed ridiculous, to somebody.

  33. Anna Phor said,

    July 14, 2009 @ 11:28 am

    "Wee-gur" seems like the least horrible solution to me.

    The sound represented by the g is a voiced velar fricative. It's the voiced correspondent to the sound represented by the ch in Scottish "loch"–which most speakers of (non-Scottish) English render as a [k]. Since [g] is the voiced correspondent of [k], makes sense to me that the ɣ would be rendered as [g].

  34. Bloix said,

    July 14, 2009 @ 11:32 am

    Vanya – take a look at fn 9 in the article linked by Bob Violence, above – the Pinyin for Uyghur appears to be "weiwuer" (diacritics omitted). So we have two Latin spellings – Uyghur and Weiwuer. Whoever is in charge of Pinyin seems to have heard a second "w," just as I did. (No, it's not just in the Mandarin, it's in both.)

    And if you read the article, you'll see that the overriding concern of the developers was to produce a transliteration system that would correspond one-to-one with the Uyghur standardized Arabic script – for purposes of computer data-entry. Ease of use by actual human readers was a distant fourth on the list of criteria. Essentially, we have a transliteration system that works great for machines and for people who are bilingual Uyghur-any European language, and sucks for anyone who is trying to read a newspaper. That may be an unfortunate necessity but it's nothing for linguists to be proud of.

  35. GAC said,

    July 14, 2009 @ 11:36 am

    On #6, your pinyin shows Ka1shi2ka1′er3 — but to me it sounds like the second ka1 is actually ge3. Is this a case of alternate transliterations or an error?

  36. Bloix said,

    July 14, 2009 @ 12:06 pm

    Franz- I suppose I started this (I say with modest pride). As I said somewhere in this mass of scribbling, if it's going to be "Uyghur," well, we can live with that – but Victor wants us to pronounce it as close to the original as possible. Then I (inadvertently? gleefully? can't quite recall my motives at this late date) stuck a stick into the beehive by complaining about the disjunction between spelling and pronunciation.

    I will say that sometimes multi-lingual, or at least language-sophisticates, seem to take pride in being able to pronounce words that whose pronunciation isn't obvious to monolingual readers. Ouagadougou? No sweat. Qatar? Rolls off the tongue. Anuradhapura, Myanmar, Xuzhou, Huehuetenango? just part of everyday conversation.

    Some of this, you can't make any easier. Foreign words are foreign, and hard for many to say. But why make it harder with odd spellings? Why does my atlas have Khmel'nyts'yy instead of Chmelnitsky? Whose interests are being served with that spelling? And if we have to spell it Uyghur, why not let people say "Y-gur," which is at least plausible from the spelling, instead of making them learn pronunciation rules for a language they barely know exists, just in order to read a newspaper article?

  37. Carl said,

    July 14, 2009 @ 12:08 pm

    Surely anyone out there who is really interested in how to pronounce these various names in their various forms can just Google and find the audio files in this post? I did, and it was very handy! :)

    And, in any case, surely it's hardly worth anyone's time worrying about whether any given romanised spelling happens to be easy for Anglophones to grasp when (romanised) English spelling itself is not a terribly good guide to English pronunciation. ;)

  38. vanya said,

    July 14, 2009 @ 12:21 pm


    That sound in Uyghur is NOT a "w", it's really not. Please listen more carefully. It's neither a "g" nor a "w", but a "ɣ", more like a "g" in the back of the throat. Now, Chinese doesn't have a "ɣ" sound any more than English does. If the Chinese choose to transcribe it with a "w" that's their choice but English has a long tradition of transcribing "ɣ" with a "gh" – "Afghan", "Baghdad", "Mossadegh", etc. Why use a Chinese phonetic spelling when we have a long established English tradition for this sound?

    "Ease of use by actual human readers " – you're not actually suggesting that Turks aren't human are you? Try to be a little more careful with your phrasing. The word "Uyghur" is not a random transcription from Arabic invented by crazed linguists – it actually maps to the Turkish latin alphabet pretty well (see http://tr.wikipedia.org/wiki/Uygur) and, as I pointed out, has been in use in English since at least 1875. I think "Uigur" goes back even further. If you want to blame someone for this word blame the British colonial service, not linguists. Maybe we could all agree on the simple French "Ouïghour".

  39. vanya said,

    July 14, 2009 @ 12:24 pm

    "And if we have to spell it Uyghur, why not let people say "Y-gur,""

    Personally I have no problem with that. That makes more sense than changing a historical spelling to reflect sounds that English speakers will probably never get right anyway.

  40. aabram said,

    July 14, 2009 @ 12:35 pm

    Bloix, English is always going to have hard time with spellings and one will never get all people to read and pronunce alike, no matter which spelling you choose. Is it really worth changing one set of dirty linens with another one? Especially in this case when some of us (me too) can clearly hear g/h and some can't?

    > And if we have to spell it Uyghur, why not let people say "Y-gur"

    And a good many people probably would sign the petition to call call Europe a Urp too, since that spelling makes more sense to them personally. What would it prove?

  41. Franz Bebop said,

    July 14, 2009 @ 12:40 pm

    Vanya, would you favor respellings of "Uyghur" as "oi-gur" or "oy-gur" ?

    Maybe it's best if respellings don't render the first syllable as "wee."

    The point here is not to represent accurate Uyghur pronunciation, but instead to find an English pronunciation that is as close as reasonably possible. "Wee" and "oy" are about all there is to choose from. Right? "Oy" also looks more similar to "Uy".

  42. Franz Bebop said,

    July 14, 2009 @ 12:42 pm

    I'm using the word "respelling" here to refer to what you folks have been calling the "pronouncer," since I'm pretty sure "respelling" is the right name for it.

  43. Bloix said,

    July 14, 2009 @ 12:47 pm

    No, of course I'm not suggesting that Turks aren't human. Are you suggesting that Turks are going to read their own language in transcription? According to the article linked by Bob Violence, the transcription protocol that was developed for Uyghur was designed to make it easy for computers to read the Latinized version – that is, a one-to-one correspondence from the Arab script that human readers of Uyghur use and a Latin version that can be input on a QWERTY keyboard. The one-to-one correspence permits easy machine transliteration. It wasn't designed for human beings to read.

  44. J. W. Brewer said,

    July 14, 2009 @ 12:58 pm

    @Bloix, since in the initial post Victor Mair expressly invited suggestions for "appropriate Americanized pronunciations," I'm not sure he's fairly accused of encouraging the faux-foreign show-offy practice which I too often find irksome. Of course, it might be the case that under ordinary circumstances (prior to the events which have put Sinkiang etc. in the non-specialist news) the only people who have frequent occasion to use the word "Uighur" in English conversation are academic specialists who can arguably use an echt-foreign pronunciation among themselves without being merely pedantic/pretentious (or at least no more ped/pret than is socially acceptable in such circles). Uighur appears outnumbered by Uygher in raw google hits (which are of course an imperfect metric for all sorts of reasons), but not so dramatically so that I'm going to invest the effort to change my own practice. But I'm a reactionary who still hasn't added the gratuitous-seeming final e to Surinam.

  45. kip said,

    July 14, 2009 @ 1:08 pm

    @vanya: "That sound in Uyghur is NOT a "w", it's really not."

    I challenge you to give 100 random American-English speakers the first pronunciation of "Uyghur" in this post (not even the Mandarin version, just the *very first* pronunciation) and ask them to write the word they hear. I'll bet nearly all will write "weewer" or "weewur". A few might write "weewher". I doubt any of them start the second syllable with a "g", unless they have some background in linguistics.

  46. acilius said,

    July 14, 2009 @ 1:27 pm

    @Language Hat:

    Very good point.

    @Franz Bebop:

    Of course, the letter in Achilles' name is a chi, not a kappa. And I certainly agree that "Akhilleus" is a ridiculous spelling. "Achilleus" may sometimes be necessary, though I can't for the life of me think when it would be.

    I think the "pronouncer" in question is the lector whose voice pronounces the names in the posted audio files.

  47. vanya said,

    July 14, 2009 @ 1:30 pm


    The English words "Uyghur" and "Kashgar" predate any transcription protocol by well over 100 years. Of course those words were designed for human beings to read. There was no one else to read them in those days.


    If you heard a live native speaker rather than a recording I doubt you'd transcribe it "w". There's a reason it's transcribed with a "g" in Russian, Japanese, Spanish, French, German, etc. And I don't have a background in linguistics – I just happened to live in Kazakhstan for many years. But even if you're right – so what? I think most of us agree that trying to accurately transcribe Uyghur phonetics into English is a waste of time. I think Bloix's suggestion that "Y-gur" be considered an acceptable pronunciation is perfectly sensible. It's no different than pronouncing "Baghdad" to rhyme with English "bag dad" or "Pakistan" to rhyme with "taxi can", both considered perfectly normal.

  48. Ellen said,

    July 14, 2009 @ 1:35 pm

    Vanya, you are missing the point of this thread. There is no such thing as "a very clear "ɣ" sound" in English. English speakers will not hear the ɣ sound in that word. Not unless they also know a language that does have that sound. Nor will they reproduce that sound. I, who do not have ɣ in my phonemic repetoire, hear the word as weewer. Others who have posted hear it that way as well.

    It seems to me that the views of those who can actually accurately hear the ɣ sound aren't relevant to the original question. Which is how to represent in writing the pronunciation for average English speakers to say the word. Which means sticking to sounds that average English speakers can hear and say. Perhaps Victor Mair invited suggestions precisely because his ear is too good and he can't hear these words like an average English speaker.

  49. vanya said,

    July 14, 2009 @ 1:46 pm


    I thought the point was to come up with a standard spelling and pronunciation for Americans. I suggest retaining "Uyghur" as the spelling since it already has a long tradition in English, and corresponds more or less to the spelling in most European languages and Turkic languages with latin alphabets. I don't expect Americans to reproduce "ɣ" so I think Bloix's suggestion of a "Y-gur" pronunciation is fine, as I've already noted twice. I think "weewer" makes absolutely no sense as a spelling, either historically or as an accurate representation of Uyghur pronunciation, so why adopt it? Again, we're writing English for English speakers, not for Uyghurs or Chinese.

  50. Ellen said,

    July 14, 2009 @ 1:53 pm

    Vanya, from the original post:

    As to what sort of "pronouncers" should be developed for the major media, presumably to represent appropriate Americanized pronunciations, I invite suggestions.

    No one suggested weewer as a spelling, nor as a representation of the Uyghur pronunciation. It was suggested as a representation of how Americans should pronounce the word, and given representation of how Americans hear the Uyghur pronunciation.

  51. David Eddyshaw said,

    July 14, 2009 @ 2:08 pm

    The spelling 'gh' in early modern English did *not* in fact represent the voiced ɣ but the corresponding voiceless sound, like the German 'ch'. Shakespeare's 'night' sounded much like modern Scots 'nicht' (or German 'nicht').

    Old English did have ɣ, written with the same letter as the stop 'g' but the sound had been lost by Middle English times.

  52. mollymooly said,

    July 14, 2009 @ 2:50 pm

    It's natural to approximate an exotic sound with what seems the closest phoneme in your native language and accent.

    The same foreign phoneme may be approximated by different phonemes in different accents of a single language. Some anglophones pronounce German "foehn" like "fern", others like "feign". Equally, the GH of "Uighur" may be /w/ for some and /g/ for others.

    Also, the English phoneme that sounds closest for an anglophone may not be the English phoneme that would sound closest for speakers of the original language.

    The pronunciation units of major broadcasters presumably have a generic policy for when a name previously known only to specialists becomes hot news. My suggested policy would be to start with the pronunciation specialists use and make concessions for exotic sounds but not for exotic spelling. Try to lead, and if the masses nevertheless adopt a spelling pronunciation, then follow them.

  53. dr pepper said,

    July 14, 2009 @ 3:14 pm

    @Franz Bebop said,

    I'm using the word "respelling" here to refer to what you folks have been calling the "pronouncer," since I'm pretty sure "respelling" is the right name for it.

    But what we seem to be looking for is a transliteration.

  54. William W said,

    July 14, 2009 @ 3:24 pm

    This is what I heard.
    Uyghur / ئۇيغۇر – [ujɣur]. My transription [ʍiɣur]

    Ürümchi / ئۈرۈمچی – [yrymtʃi]

    Täklimakan / تەكلىماكان – [tæklɨmakan]. My transription [tæçlɨmakan]

    Tarim / تارىم – [tarɨm]. My transription [tʰɑrɨm]

    Tarim Daryasi / تارىم دەرياسى – [tarɨm daryasi]. My transription [tʰɑrɨm dərjɑsi]

    Kashgar / كاشگار -[kaʃgar] or Qäshqär / قەشقەر – [qæʃqær]. My transription [kəʃGər]

    Tängri Tagh / تەڭرىتاغ – [tæŋri taɣ]. My transription [təŋrɨ təx]

  55. Simon said,

    July 14, 2009 @ 4:17 pm

    In the word of copy editing (or copyediting, if you follow AHD instead of MW), which I concede is not a world fans of Language Log are necessarily open to, issues of "correct" spelling or pronunciation are resolved by following the primary spelling/pronunciation in your dictionary of record and your style guide of record. In the publishing industry, these are typically Merriam Webster and the Chicago Manual of Style (thought not always). Merriam Webster uses the spelling Uighur, and their pronunciation is given as \ˈwē-ˌgu̇r\, which we might IPA-ify as something like [ˈwi:gɹ] for rhotic speakers or [ˈwi:gɝ] for nonrhotic. This is not to say these are "more correct" than others; only that the dictionary through its lexicographical methodology has identified these as reflecting widespread, educated usage.

    The reason why writers resort to references of record is to ensure both clarity and consistency (so that multiple writers can be using the same spellings). The dictionary pronunciation also gives you a guide to what you might expect educated native speakers to use as a pronunciation. If you use another pronunciation, even if "more authentic," you *might* increase risk of being misunderstood, depending on your target audience.

    Anyway, when it doubt, it's more than acceptable and quite professional, as a writer, to defer to your preferred dictionary and style guide of record for questions like these.

  56. language hat said,

    July 14, 2009 @ 7:15 pm

    I heartily second Simon's comment.

  57. Spectre-7 said,

    July 14, 2009 @ 7:39 pm

    After far too many listenings, I'm tempted to partially agree with both sides as to how Uyghur is pronounced in the sample. The problem, I think, is that both sounds are present, but that different listeners are likely to focus on one to the exclusion of the other.

    My IPA is, as always, trash… but what I hear is something like ujɣwur. Of course, I'd be curious to listen more samples, preferably with video.

  58. grackle said,

    July 14, 2009 @ 9:05 pm

    It sounds to me like the speaker in the sound samples is a native speaker of a Chinese not a Turkic language. Might this be a factor in the disappearance of the غ sound? The Turkic languages of central Asia mostly used the Arabic script until replaced in the Soviet era by Cyrillic. Most European transcription systems for the Arabic/ Persian script have used ‘gh’ for غ since the late 19th century. I don’t know the extent to which Uyghur has a sound that corresponds to the sound of a true غ . I suspect that that there is a faint guttural suggestion, that might be interpreted as a soft 'g' sound to an English speaker. At any rate a true غ is notoriously difficult for a non-native speaker to reproduce, hence the possibility of its evidence in the sound sample being slurred over.

  59. Andrew M said,

    July 15, 2009 @ 3:03 am

    Perhaps Bloix's point could be made clearer in the opposite direction, even if not in a purely parallel fashion. Imagine a(n?) Uyghur-speaker is transliterating the word Internet into Uyghur based on the recording of a word. Only, the recording was made by an American, and in standard American English speech, the word Internet is pronounced with the first t elided, and the second either unreleased or glottalized. In essence, speakers of Uyghur might not hear any t's at all, and they might just find a transliteration of the word containing the equivalent of t to be non-sensical.

  60. mollymooly said,

    July 15, 2009 @ 6:49 am

    On Google news, I'm getting 4,710 hits for "a Uighur" and 246 for "an Uighur".

  61. Ken Brown said,

    July 15, 2009 @ 8:09 am

    I'm afraid I'm one of those that can't hear a "w" in there. In the Chinese version yes, but the supposedly Uighur version sounds like a "gh" to me. If we are not allowed that then "g" sounds nearer to me. Maybe I say "g" differently from how many Americans might say it.

    If we are looking for a better transliteration of the word for an English speaker, and if "gh" is no longer understood, and if "g" sounds wrong to speakers of the language, then maybe "weehur"? The "h" has no clue it should be voiced though so it can;t be quite right, but its surely nearer than "w"?

    I would prefer "weeghur" or even "weekhur" (the first consonant in words like "Khan" might be familiar enough to English speakers for them to get it right in a Central Asian context).

    However I'm not a linguist, I speak no Turkish or Turkic of any sort and I've never been to Central Asia. So I don't know what it really sounds like

    But why all this concetrtation on the "gh"? The consonant that I can't say is the "r" at the end. It is not at all close to anything that I would be likely to use when talking English and when I try it I come out with a sort of unpleasant grunt.

    To me all the Uighur pronounciations in those clips sound pretty much the way I'd expect them to sound from the way they are spelled. Its nice to no I've been hearing "Kashgar" correctly in my mind all these years. Even though its not a word that I might ever have actually said out loud.

  62. Karen said,

    July 15, 2009 @ 9:21 am

    I'm using the word "respelling" here to refer to what you folks have been calling the "pronouncer," since I'm pretty sure "respelling" is the right name for it.

    But what we seem to be looking for is a transliteration.

    Actually, what he asked for was indeed "pronouncers" – that is, guides for broadcasters so they know how to pronounce the word, NOT a "respelling" or a "new transliteration". It's like adding (MADE-off) to a story about Bernie Madoff – nobody's suggesting we "respell" his surname.

  63. Karen said,

    July 15, 2009 @ 9:21 am

    err, "guides for broadcasters and readers so they know how to pronounce the word"

  64. William W said,

    July 15, 2009 @ 9:29 am

    Here are my revised transcriptions.

    About the word "Uyghur", this time I heard a labialized sound for the middle consonant and have transcribed it this way: [wiɣʷur]. This might explain the presence of both the 'gh' in the common Anglicized form and the perception of some [w] like sound.

    Uyghur / ئۇيغۇر – [wiɣʷur]
    Täklimakan / تەكلىماكان – [tʰεçlɨməkən]
    Ürümchi / ئۈرۈمچی – [yrymtʃi]
    Tarim / تارىم – [tʰɑrɨm]
    Kashgar / كاشگار – [cεʃGər]
    Tarim Daryasi / تارىم دەرياسى – [tʰɑrɨm dərjɑsi]
    Tängri Tagh / تەڭرىتاغ – [tæŋrɨ tɑx]

    Andrew M's point about the disparities between the written form and an allophonic transcription are on the mark I think. Especially with a number of the vowels, I suspected that perceived differences were due to the environment of the sound and were allophonic rather than phonemic.

  65. Nathan said,

    July 15, 2009 @ 9:46 am

    It's my understanding that the "exotic" consonant in question is a uvular [ʁ] rather than a velar [ɣ].

  66. vanya said,

    July 15, 2009 @ 10:18 am

    It sounds to me like the speaker in the sound samples is a native speaker of a Chinese not a Turkic language.

    Actually I had that same impression, but it may be because I'm subconsciously listening for a Kazakh accent and this speaker has a much "softer" accent. In Kazakh or Kyrgyz the "gh" in Uyghur would not be mistaken for a "w".

  67. J. W. Brewer said,

    July 15, 2009 @ 11:56 am

    In hopes of getting a sense of English usage undistorted by modern political concerns (although perhaps distorted by bygone political concerns), I took a look at the index of a recent paperback reprint of Byron's The Road to Oxiana (first published 1937), which is in my bought-but-not-yet-read pile. The author unfortunately seems not to have referred to the U-people (or their language) under discussion here, although he uses the pleasingly archaic spelling "Uzbeg" for Uzbek. He apparently had hoped to make it to Kashgar (thus spelled) but never did. He does describe what Dr. Mair neutrally refers to as "East Central Asia" as variously "Chinese Turkestan" and "Sinkiang."

    The latter was the most common name and spelling used for the area in the various English-language atlases I was exposed to as a boy in the 1970's when my own sense of standard spelling was being formed and solidified. Accordingly, it is my view that it should remain the normative English spelling for non-specialist contexts since any supposed advantages offered by Pinyin are not worth the transition costs. Moreover, it is my sense that the particular romanization involved (the "Postal Map" system) was one where ease-of-use-by-foreigners was given considerable weight as against the various other competing legitimate desiderata for a romanization system, and is therefore especially suitable for answering the "what should we call it in English?" question.

  68. Roderick said,

    July 15, 2009 @ 3:02 pm

    I hesitate to break into this elegant discussion, but not being monolingual (English and French since earliest childhood) I do hear a faint g in the For interest, Aurel Stein (Serindia, Oxford 1921, vol.III, p.1130) consistently uses the spelling Uigur, and notes that Kao Chü-hui (Gao Juhui in pinyin), writing in A.D. 938, refers to the Uigurs using the Chinese characters Hui-ho (回紇 Huihe). At that time, the Uigurs were in power in Ganzhou (now Zhangye). The modern pocket dictionary Xinhua zidian also gives this transcription Huihe, noting that the second character was later changed from 紇 he to 鶻 hu.

  69. Roderick said,

    July 15, 2009 @ 3:10 pm

    …I do hear a faint g in the first "pronouncer". As for "whoever is in charge of pinyin" this system simply transcribes the Chinese characters now used for the Uigur nationality, namely 維吾爾 wei wu er, rather less elegant that the Tang version, Hui he.

  70. John Swindle said,

    July 16, 2009 @ 3:30 am

    GAC said about the Mandarin version of example #6, "it sounds like the second ka1 is actually ge3. Is this a case of alternate transliterations or an error?"

    And the final syllable as heard in the clip is not er3 but er4. Aren't there different ways to write the long version of Kashgar in Chinese?

  71. Pronouncing Uyghur proper nouns said,

    July 16, 2009 @ 9:05 am

    […] few days ago on Language Log, Victor Mair wrote an excellent post on how to pronounce various proper nouns in Uyghur, the Turkic language spoken by the Central Asian people native to northwest China. Following the […]

  72. Franz Bebop said,

    July 16, 2009 @ 9:23 am

    After reading all of the comments, I'm wondering what the original problem was with the respelling provided New York Times ("WEE-gurs"), which was Prof. Mair's original reason for posting.

    Bloix's main complaint was that it seemed strange to be asked to pronounce "gh" as a w-sound. However, numerous people on the thread pointed out that the original ɣ sound in Uyghur isn't like an English w-sound at all, even though a English-speaker might mistakenly think so. What's more, it might come across as insulting to use a w-sound in the anglicized pronunciation, since that seems to follow Mandarin rather than Uyghur. Fair enough.

    So, then, what is the problem with the respelling "WEE-gur" offered as a pronunciation guide for English speakers? It seems fine to me. What are the alternatives?

  73. Ben said,

    July 16, 2009 @ 10:53 am

    Am I the only one who hears the /kl/ in Taklimakan as a lateral fricative? It's my favorite phoneme; all the cool languages have it: Greenlandic, Welsh, Zulu, Mongolian, and Navajo, to name a few.

  74. Linguistic Data » Blog Archive » “Xinjiang Proper Noun Primer” said,

    July 16, 2009 @ 1:52 pm

    […] Source URL: http://languagelog.ldc.upenn.edu/nll/?p=1576 […]

  75. Henry said,

    July 16, 2009 @ 7:32 pm

    Ben, I also thought the /kl/ sounded like a lateral fricative.

    [Add to the list of cool languages: Tlingit (where the /tl/ is a lateral fricative, spelled "L" in Tlingit orthography so the language is "Lingit")]

  76. john riemann soong said,

    July 17, 2009 @ 4:10 am

    I have a feeling that the "chi" in "Urumchi" is closer to the Mandarin "qi" than the English "chee". But I could be wrong.

  77. Bob Violence said,

    July 17, 2009 @ 12:30 pm

    The Uyghur "ch" is tʃ, which is identical to the hard "ch" in English. The Mandarin "q" is tɕʰ.

  78. Victor Mair said,

    July 17, 2009 @ 3:51 pm

    Sent to me by a friend who wishes to remain anonymous.


    Your blog has:

    I don't know if 'combination of Turkic and pinyin' is a useful formulation. It would be better to call it, 'a defunct latinisation' or' the Uyghur latinisation (New Script) of 1969(?) — 1983', or something of that sort.

    Let us look at the more-or-less remarkable letters found in the defunct script:

    Q X : for IPA [ʧ ʃ ] ) ; letters whose near values are taken over from Hanyu Pinyin
    Ə Ƣ Ɵ : for IPA [æ ɣ ø ]; letters taken over from USSR latinisations (and retained in later cyrillicisations)
    Ⱪ Ⱨ Ⱬ : for IPA [ q h ʒ ]; new letters. The Ⱪ is no doubt based on Cyrillic Қ, with the other two by analogy

    All in all, only two letters from Hanyu Pinyin.

    Unrelated, it seems to me that in the above you ought to have written 'Hanyu Pinyin', not simply 'pinyin'; and, if simply 'pinyin', then it should be 'Pinyin' — in accordance with the usual treatment for proper nouns in English. I'm sure that you will agree that in English (where the word refers exclusively to Hanyu Pinyin), unlike in Chinese (where it refers to any 'pinyin wenzi' or ' pinyin fang'an '), it is without question a proper noun.

    But it's not only a matter of transliterative fidelity to the Perso-Arabic. Consider:

    The word 'in Uyghur' ('in the Uyghur language'):

    defunct latinisation : Uyƣurqə

    USY: Уйғурчә

    UEY : ئۇيغۇرچە

    ULY: Uyghurche

    In which we see: not i but y; not ى but ي ; not и but й ; not i but y .

    (Not just UEY but USY, and ULY are in use. Radio Free Asia for example has any Uyghur language page available in all three scripts. Likewise the Qur'an, and I believe some Christian material, is available in all three scripts.)


    VHM: Key to the different Uyghur transliterations cited above:

    UEY: Uyghur Ereb Yéziqi (Arabic Uyghur — post 1982 spelling)

    UKY/ULY: Uyghur Kompyutér Yéziqi (Romanised Uyghur—originally developed for computer use in 2001, increasingly recognised as the new standard romanisation. It is also called ULY — Uyghur Latin Yéziqi since 2004.)

    USY: Uyghur Slawyan Yéziqi (Cyrillic Uyghur, used in the former Soviet Union).

  79. Victor Mair said,

    July 18, 2009 @ 8:49 pm

    Just to clear up a couple of loose ends and draw some sort of conclusion, first, John Swindle and GAC ask if there are different Chinese transcriptions of Kashgar, and the answer is yes. There are 喀什喀尔, 喀什葛尔 (note that because of sandhi, the penultimate 3rd tone becomes 2nd tone in front of the final 3rd tone), 喀什葛儿, etc., so this accounts for the different pronunciations that one hears for the name of the city (and the river) in Mandarin. Usually, however, the name is just chopped down to 喀什 Kashi.

    Finally, before 788, the Uyghurs were referred to in Sinitic as YUAN2HE2 袁紇 (Early Middle Sinitic [EMS, 601 AD] wuan ɣɘt) and then WEI2HE2 韋紇 (EMS wuj ɣɘt). From 788 on, the Uyghurs came to be called HUI2HE2 回/迴紇 (Late Middle Sinitic [LMS, mid-8th c.] xɦuaj xɦɘt) and then HUI2HU2 回/迴鶻 (LMS xɦuaj xɦut).



    1. 紇 has the variant reading GE1 as in GE1DA ("knot"), but this second syllable in the name of the Uyghurs is conventionally read as HE2.

    2. 鶻 can also be read as GU3, but in the name of the Uyghurs it is conventionally read as HU2.

    3. I suspect that the final -t in all of these transcriptions is standing in for final -r, which was lacking in Sinitic at the time.


    Centuries later, a quite different transcription became popular, namely WEI4WU2ER2 畏吾兒, ultimately leading to the modern Mandarin form WEI2WU2ER3 維吾爾.

    What does this all add up to? As I just concluded in my latest comment to the other Uyghur thread (http://languagelog.ldc.upenn.edu/nll/?p=1579), the medial consonant of Uyghur seems to have been interpreted differently by Sinitic speakers at various times and in various places. What is interesting and ironic, in the light of our very lengthy discussion in these two threads, is that — whereas English speakers seem to be settling on /g/ for what they hear or accept by consensus as the medial of the name — Mandarin speakers have unambiguously accepted /w/. If we add the English and Mandarin together, we come close to what that most elusive Turkic sound actually is, a fact that many astute commenters to these two threads have already pointed out.

  80. Gardner Bovingdon said,

    July 20, 2009 @ 10:46 am

    I have read this thread beginning to end and found it fascinating. It seems to me there are two very different constituencies weighing in here: one is the group that thinks Romanizations for English speakers should either (a) accurately capture the sounds of a foreign language for those who do not speak that language or (b) render the word(s) in a way that can be pronounced by, and will not seem idiosyncratic to, said English speakers – without trying to be faithful to the original. Some people may have noticed that English has, um, absorbed a few foreign words over time, and also that speakers of English don't much trouble themselves to follow the pronunciations from the original languages. One poster actually mocked the attempt to hew to original pronunciations as a " faux-foreign show-offy practice" – and I think that captures the view of many Americans. So, English speakers should say Wee-grr. That is standard practice when the term is used by specialists in an Anglophone setting. But those English speakers should accept the spelling "Uyghur," not because it's particularly easy to remember or makes intuitive sense, but because _that's how Uyghurs wish the name of their language and culture-group to be spelled_. There's a political issue here that has not been touched on as far as I can see: Uyghurs have mostly had their language, their culture, their ethnonym, and their history represented to the world by others, for the most part. It's time for that to stop. And that is why, as several people have pointed out, it's important not to assimilate the spelling or pronunciation to the Hanyu Pinyin versions.

    As for the linguists' camp, I assume that multilingual folks are ready to accept the obvious fact of diversity of pronunciation _within the community of native speakers_. Ask someone from Seattle, someone from Boston, and someone from Mobile how to say "car" and then tell us, which is the correct pronunciation? Similarly, there are Uyghurs who pronounce the "gh" (I'm too tired to hunt up the gamma) in the name very lightly, so that to the untrained ear it might sound like a "w," a "wh," or an "oo". Others pronounce it with strong emphasis, so that it is almost a "g." And still other Uyghurs (here I may draw some ire) scramble the innards of the name, saying "Urgooey" (all of these, by the way, are non-linguist, non-IPA, English-speaker-friendly Romanizations I've conjured on the spot).

  81. Ellen said,

    July 23, 2009 @ 10:50 am

    Gardner, there are also those of us who weren't talking about Romanizations at all, but rather about pronunciation guides, which was the original subject. Although, come to think of that, there's probably the same two groups with that.

  82. David Marjanović said,

    August 8, 2009 @ 8:48 am

    Contrary to my expectations, the first syllable of Uyghur really is closer to [wi] than to [uj]. It doesn't start with a glottal stop or anything. I wonder if interference from the Mandarin wu (which does not even start with a real [w]!) is involved.

    Also, likewise contrary to my expectations, it really is [ɣ] (first time) and arguably [ɣʷ] (second time), not [ʁ]. However, I can hear the friction very clearly; it's not an approximant like [w].

    In Ürümchi, the speaker uses (apart from tone) the exact same [t͡ɕi] for both languages. It's neither aspirated as I'd have expected for Mandarin (…let alone the aspirated dorso-palatal affricate of Standard Mandarin), nor is it [t͡ʃ] as I'd have (perhaps too naïvely) expected for any Turkic language.


  83. The Uighurs, China’s Embattled Muslim Minority, Are Still Seeking an Identity | TokNok Multi Social Blogging Solutions said,

    September 27, 2013 @ 12:05 pm

    […] from the far-western province of Xinjiang. Like many minorities, the Uighur, the native people of Xinjiang, have made their chief impact on mainstream culture through cuisine. I have always favored their […]

  84. The Uighurs, China’s Embattled Muslim Minority, Are Still Seeking an Identity | The Daily News Report said,

    September 27, 2013 @ 12:08 pm

    […] from the far-western province of Xinjiang. Like many minorities, the Uighur, the native people of Xinjiang, have made their chief impact on mainstream culture through cuisine. I have always favored their […]

RSS feed for comments on this post