Uyghur, Cantonese, and other valuable languages of China

« previous post | next post »

In the Sinosphere section of yesterday's NYT, there's a thought-provoking article by Didi Kirsten Tatlow titled "Speak Uighur? Have Good Vision? China’s Security Services Want You" (2/19/16).

"Question of characters" (5/20/06)

"Teach simplified characters" (8/31/94)

[N.B.:  The South China Morning Post (SCMP) is behind a paywall, so you may have difficulty directly accessing some of these articles.]

If Cantonese speakers wish to gain some respect for their language, they themselves should stop referring to it as a "dialect" (or "only a dialect") in English, for that is a mistranslation of the Chinese term fāngyán 方言, which should more precisely be rendered as "topolect".

Google on "victor mair language log intelligibility" and "victor mair language log dialect".

As to how to render the English word "dialect" in Chinese, I would propose the neologism "tōngyán 通言".

As for the

[Thanks to Robert S. Bauer]


  1. Gene Anderson said,

    February 20, 2016 @ 3:14 pm

    Right on. It really annoys me to hear Cantonese or Hokkien called "dialects." A lot of people refer to Native American languages as "dialects" on purely racist grounds. One recalls the rather exaggerated comment on Portuguese that a language is (sometimes) only a dialect with a navy….

  2. Bathrobe said,

    February 20, 2016 @ 3:43 pm

    The article also mentions: In the post, a person named Wang Yuele, which can also be transliterated as Wang Yueyue".

    While this is arguably a kind of shorthand, it's a misleading one. The process is not one of 'transliteration'; it's one of figuring out the person's name, or more specifically how it is spoken. The person is either Wang Yuele or Wang Yueyue, and only the person him/herself would know that. It's a similar dilemma to that of whether 'Reagan' should be pronounced 'Reegun' or 'Raygun', and that was resolved when it became known that Ronald Reagan pronounced it 'Raygun'. (This was an issue in the Japanese press, which initially wrote Reagan's name リーガン riigan but later changed to レーガン reegan).

    On the other hand, I'm not sure what kind of wording would make this clear. Wang Yuele, which can also be read as Wang Yueyue is the traditional wording, but I suspect that only those familiar with Sinographs would understand this concept. Wang Yuele, which can also be pronounced Wang Yueyue might be better. But whichever is chosen, 'transliteration' is the wrong concept, because it implies that the choice between 乐 as yue or le is equivalent to that between Достое́вский as Dostoyevsky or Dostoevsky. It's not. There's no difference in the Russian pronunciation of Достое́вский; the only problem is how it's rendered in Roman letters.

  3. leoboiko said,

    February 20, 2016 @ 5:09 pm

    Where can I learn more about syntactic differences between Chinese languages (including Literary Sinitic)?

  4. David Morris said,

    February 20, 2016 @ 7:13 pm

    Ah yes, I thought so, 'This could reflect an interest in the online monitoring of citizens and in the observing of China’s ethnic minorities'. 'Could'???? Why else? I doubt if the security services are doing this just to be nice to speakers of those languages.

  5. Victor Mair said,

    February 20, 2016 @ 8:27 pm

    From Didi Kirsten Tatlow:

    When I started out as a journalist, in Hong Kong in 1994, I would try to talk to my fellow HK reporters in Chinese. They invariably answered in English. One day, frustrated, I asked one why he did that. He said: "Because if I don't I'm afraid you will think I don't speak English."

  6. liuyao said,

    February 20, 2016 @ 9:33 pm

    I suspect the student's complaint was the use of "Cantonese characters" (many with the mouth radical and purely phonetic), instead of traditional characters, which as Alex Lo remarked would indeed be absurd. Being from a student organization, it's likely to have more of these vernacular characters than official documents from the university.

    Don't know where Wang Yuele is from, but she may pronounce it in neither of the two in her mother tongue (many yue are often pronouced yao in many regions). I agree "transliteration" is misleading; they could simply say "rendered." I'm hesitant though to let everyone render their name by a pinyin proxy to the local pronunciation. Whichever transliteration, the media (on the web) should start appending the characters (at first mention), otherwise it's almost impossible to find more about the person (unless that's the intent).

    To make readers of this blog less angry with the mention of dialect, take solace in other translations that are equally inaccurate, such as empire, dynasty, (Middle) Kingdom, dragon, phoenix. To take it off from the journalist or editor, was Herbert Giles (recently featured here) not aware of the linguistic diversity of China, or of the dialects of England? [Just to be clear, I of course welcome your neologism "topolect" to be more widely used. I started using it immediately after first saw it.]

  7. liuyao said,

    February 20, 2016 @ 9:52 pm

    While coming up with inept translations, there was an old term that I believe, albeit inaccurate if taken its European meaning, is much better than the new translation: "canonized" for (being given) shi 謚, which is now rendered "posthumous name."

  8. Jeff W said,

    February 21, 2016 @ 6:16 am

    Alex Lo and other influential intellectuals could help clarify matters greatly if they would stop thinking that there's something out there called "Chinese"…

    Isn’t Lo referring to Article 9 of the Basic Law which he quotes and which says “In addition to the Chinese language,…”?

    I think what he is saying is whatever that refers to “can be spoken in Putonghua or Cantonese and written in simplified or traditional scripts.” Maybe he’s wrong and the specific reference to “the Chinese language” refers to Cantonese or traditional scripts or both only but I don’t think he can be faulted for what is a reference in the Basic Law. (I’m actually surprised I am defending Alex Lo.)

  9. Victor Mair said,

    February 21, 2016 @ 8:03 am

    What is needed is for intelligent people in Hong Kong and elsewhere to point out the murkiness of conceptions of "Chinese" such as that in the Basic Law. So far as I know, no one (except me) has ever done so.

  10. K Chang said,

    February 21, 2016 @ 1:12 pm

    Kinda makes you wonder if they need undercover operatives… or just someone to teach AI pattern recognizers for the great firewall of China (and whatever their NSA's equivalent of project Carnivore, the massive surveillance project).

    Speaking a relatively obscure dialect is a valid tool on opsec (operation security) as it's almost as good as speaking in code.

    Which lead me to think of a novel I read recently, Darknet by Matthew Mather, which postulated that AI is already good enough to act in a person's proxy… permanently, by assuming the person's voice print and video footage (all constructed, of course) either with a person's permission (as a true proxy) for screening calls and/or attending meetings, or for darker purposes (in the novel, by contracting murders and sabotage on the darknet on competitors, buying off politicians and other influencers). In the book, the AI was able to imitate almost anyone over the phone, with the proper voice and inflections and even speech patterns (if given enough speech samples).

    I know, not related, but kind makes you wonder.

  11. Mo said,

    February 21, 2016 @ 5:26 pm

    When can a 方言 be called a language? Mutual intelligibility?

  12. leoboiko said,

    February 21, 2016 @ 5:48 pm

    @Mo: That's the only non-political, neutral criterion. However, intelligibility is gradual, hard to measure, and partially subjective; so there's no deterministic algorithm to distinguish dialects and languages. That's one reason why "topolect" is a good neologism; it allows us to discuss Sinitic varieties while sidestepping identity problems. Many traditional distinctions seem little justified; for example, I, a Portuguese speaker, can understand most of a Spanish pop song without any previous study, and yet they're called "different languages"; a Mandarin speaker wouldn't make heads or tails of a Cantonese song, and yet they're often called "dialects". It's evident that those classifications have more to do with politics and history than with linguistic fact.

    Moreover, calling a form of speech a "dialect" often goes hand-in-hand with disparaging some part of society and making them submit to the dominant class, whose speech is invariably crowned as the standard, proper, beautiful form of "the language". So linguists tend to get doubly annoyed when a mutually unintelligible variant is called "dialect"; not only it obscures scientific fact, so that distinct languages are seen through the lens of the dominant one, but it also easily lends itself to discrimination, understudy, language death etc.

  13. Neil Dolinger said,

    February 21, 2016 @ 9:05 pm

    Like leoboiko, I'd also be interested in references to comparative grammars of the different Chinese languages. Wikipedia only goes so far.

  14. Victor Mair said,

    February 21, 2016 @ 9:54 pm

    Listing only items in English, there are many useful references here, starting with Anne Yue-Hashimoto, Comparative Chinese Dialectal Grammar — Handbook for Investigators (Paris: Ecole des Hautes Etudes en Sciences Sociales, Centre de Recherches Linguistiques sur l'Asie Orientale, 1993).

    For Cantonese, I highly recommend the works of Stephen Matthews and Virginia Yip.

    For Wu, look at the publications of Richard VanNess Simmons.

    For Min, consult the research of David Prager Branner.

    It is worth noting that, except for Matthews and Yip (both with their Ph.D.'s from USC), the other three above are closely linked to the University of Washington, where the formidable Jerry Norman (a specialist on Min) was based for decades before his untimely passing in 2012.

    I think that Lisa Lai-Shen Cheng (Ph.D. from MIT) at Leiden does some explicit comparison of syntax in Chinese languages, but I'm not familiar with her publications.

    As you can probably glean from what I've written above, what you're looking for (information about comparative syntax of Chinese topolects) will not be handed to you on a plate. You will have to go digging for it on your own.

    BTW, I wanted to thank leoboiko for his reply to Mo. It is one of the clearest, most balanced statements on the problem of "dialect" vs. "language" that I have seen.

  15. Mo said,

    February 22, 2016 @ 7:15 am

    I think if people are forced to put up “请讲普通话” signs then that's indication enough of mutual intelligibility.

  16. Victor Mair said,

    February 22, 2016 @ 7:55 am


    I think that, in both of your comments, for "mutual intelligibility" you meant "mutual unintelligibility".

  17. liuyao said,

    February 22, 2016 @ 11:01 am

    I don't think we need more proof of mutual unintelligibility when it comes to (spoken) Cantonese and Mandarin. The more interesting question is that, why didn't Hong Kong, away from the influence of the central governments, develop a full-fledged vernacular literature, and instead when it comes to official and formal writings, one could point to it and say it's not Cantonese but Modern Standard Mandarin? That they themselves would think (or used to at least) it vulgar to write the vernacular characters down? I read that those characters are not permitted on school tests (essay composition?). Of course now they are all over social media and on text messages, along with the Roman letters.

    Could it be that using MSM to refer to both the spoken Mandarin and the written baihua, is just a continuation of the confusion of the term Chinese? Again it may come down to translating the terms 現代漢語, 白話文, 書面語, none of which refer to the Beijing speech explicitly. If it's ultimately about how Hong Kong people use their language(s), it's helpful to know the terms they use, instead of debating only in English.

  18. Anonymous Coward said,

    February 22, 2016 @ 12:38 pm

    High-register, formal Mandarin and Cantonese are extremely similar to each other. (Nor is really late (early 20th century, non-belletristic) Literary Chinese or formal written Japanese that different, either.)

  19. liuyao said,

    February 22, 2016 @ 12:51 pm

    Prof. Mair challenged: Show me what this thing called "Chinese" is that — when spoken — can subsume "Mandarin" and "Cantonese".

    I went to the Chinese (中文) version of scmp, and was able to understand all the headlines. I assume that Cantonese-speakers could read them equally well.

  20. Victor Mair said,

    February 22, 2016 @ 2:00 pm


    When you speak it, you're not speaking Cantonese.

  21. Eidolon said,

    February 22, 2016 @ 2:29 pm

    Full-blown spoken Cantonese is rarely written down, simply because an universally agreed-upon orthography for it does not exist. Sites such as are mainly writing Cantonese-style Standard Mandarin and the language used *could* be considered "a written dialect of Standard Mandarin" by virtue of mutual intelligibility with written Standard Mandarin, although the spoken form would still be mutually unintelligible with spoken Standard Mandarin due to pronunciation differences. But locally spoken Cantonese isn't mutually intelligible with either Standard Mandarin or Cantonese-style Standard Mandarin, though I imagine due to decades of Standard Mandarin education, Cantonese speakers, especially in Guangdong as opposed to Hong Kong, should be able to understand it.

    The goal of the Chinese government is to ultimately diminish/relegate Cantonese, Hakka, Wu, etc. to second language status. This has been obvious to me for a while. Xi's government in particular has been pushing linguistic and cultural homogenization. It is seen as a necessary step to nation-building, producing an end result similar to those of other Northeast Asian states such as Japan and Korea, which are seen as the "proper" example of a modern nation-state as opposed to the now defunct Soviet Union with its "nationalities" mess and regionalisms that the PRC tried to copy originally. Recent articles about PRC policy towards minorities such as the Uyghurs, Tibetans, etc. have all observed this trend, but it should be mentioned that the trend applies to groups that don't speak Standard Mandarin in general, not just as it pertains to languages outside of Sinitic, and this is in fully display here as Fujianese, Hakka, Cantonese or Wu are put alongside Uyghur, Mongolian, Tibetan, etc. in terms of recruitment priorities. All groups that don't speak Standard Mandarin are targeted.

  22. liuyao said,

    February 22, 2016 @ 2:37 pm

    What is it, then, that when a Hong Kong native reads the newspaper headline, either aloud or in his/her mind? MSM with Cantonese pronunciation? What is it that I'm reading? Not Mandarin as I would use in a real conversation. What is it that meets our eyes on the paper or computer screen? Chinese-writing-that-is-not-any-language-or-dialect-when-spoken.

    By the way, Cantonese 粵語 does appear on a par with English 英語, so I'm not sure how insisting on the correct English terminology helps them see their own language issues better.

  23. Eidolon said,

    February 22, 2016 @ 3:05 pm

    "What is it, then, that when a Hong Kong native reads the newspaper headline, either aloud or in his/her mind? MSM with Cantonese pronunciation?"

    Yes, unless the person is fully fluent in spoken Standard Mandarin, in which case it'd just be Standard Mandarin with Standard Mandarin pronunciation. I'd even argue that this language exists among Cantonese speakers as a third language – between spoken Cantonese and spoken Standard Mandarin, that could be used in certain situations before the speaker obtains fluency in Standard Mandarin.

    Fluency, after all, also exists within the spectrum of intelligibility. A Cantonese speaker could speak Standard Mandarin with a "Cantonese accent," in which case it'd just be a dialect of Standard Mandarin, or speak Standard Mandarin with Cantonese pronunciations, in which case it'd be a mutually unintelligible third language. Outside of all this is spoken Cantonese itself, which is different not only at the level of pronunciation, but can also be different at the level of lexicon, syntax, etc. But, as I've argued in the past, people should be aware of the fact that what people speak is constantly changing, and even local forms of spoken Cantonese today are shifting towards Standard Mandarin in many respects.

  24. Victor Mair said,

    February 22, 2016 @ 3:38 pm

    Eidolon has put the matter succinctly and clearly.

    BTW, English is part of the official mix of what constitutes "language" in Hong Kong, so it has to be taken into account too.

  25. Bob Ladd said,

    February 22, 2016 @ 4:38 pm

    I don't disagree with Victor Mair's description of leoboiko's discussion of language and dialect as clear and balanced. But I think leoboiko's idea that you can discuss "scientific fact" about this question while "sidestepping identity problems" is an illusion, because identity is central to the whole cluster of issues. It seems clear that there is no purely linguistic "scientific fact" that will allow us to distinguish related languages from dialects of the same language. Mutual intelligibility is one criterion, but identity (and all that goes with it, including political boundaries, literary traditions, and much more) is another. That is why I believe that Victor Mair is right to say that Cantonese and MSM are different languages, but also that some of the people he criticises are not wrong to continue talking about "Chinese". For a variety of historical reasons, Chinese and Arabic are fairly extreme cases in which a lot of mutually unintelligible ways of speaking are lumped together as a single "language". That doesn't mean that lumping them together is merely a sign of ignorance.

    Also, some groups of language users are "lumpers", willing to treat mutually unintelligible varieties as instances of the same language; others are "splitters", insisting that even fairly mutually intelligible varieties are different languages. (This can change, too: until 30 years ago, "Serbo-Croatian" was "a language"; now there are lots of people who would insist that there's no such thing as Serbo-Croatian.) The same applies to writing systems: it's perfectly normal to talk about "the Roman alphabet" even though that term includes a lot of mediaeval writing styles that nobody today can read without some instruction; in India, by contrast, closely similar offshoots of Sanskrit nagari count as different scripts (Hindi, Bengali, Punjabi, etc.).
    I think it's worth keeping all this in mind in trying to think straight about "Chinese" and Chinese topolects.

  26. leoboiko said,

    February 22, 2016 @ 5:27 pm

    @Bob Ladd: Two points. The first is that, by "sidestepping" the issue, I mean that there may be unclear cases. I don't think anyone will dispute that Cantonese or Taiwanese are different languages than Mandarin, by the linguistic criteria of intelligibility; and that Cantonese-style Mandarin or Taiwanese-style Mandarin are dialects of Mandarin. But there may be varieties that are closer to Putonghua-Mandarin than Cantonese, and yet farther than Taiwanese-Mandarin. Whether they're mutually intelligible is an interesting question, but may be pending better arguments. What's more, some other audience might feel, as you do, that intelligibility just isn't a valid criterion, and that, for example, being a cultural part of the nation of China is more important. With "topolect" we can just let the question open; we can talk about the syntax or morphology or history of a variant without yet having a clear case for language-ness, while also avoiding the trap of wrongly deeming it a simple dialect. It's useful to have a general term for "dialects, languages, or any other variants".

    The second is that, if I'm reading your comment correctly, you assert that mutual unintelligibility is not a sufficient criterion for language identity, because identity also depends on tradition, politics and much more. What I am saying is that it's desirable to redefine what it counts as "different language" purely by intelligibility, notwithstanding historical or popular usage (just like physicists distinguish "mass" and "weight", even though in everyday language they're interchangeable). Why is a technical definition of "language" desirable? Two reasons. First, to call something like Cantonese a "dialect" is an act that has often been used as a tool of oppression; by saying "no, according to linguistic fact (=intelligibility) it's a different language", we present a good argument to help Cantonese speakers to defend their identity. The second is that calling something a "dialect" has caused bias in language studies. The description of Japonic languages often suffered because they're culturally considered to be just "dialects" (方言) of Japanese, and therefore they were described from a Japanese-biased point of view; fieldworkers would focus on superficial lexical differences and gloss over each language's particular grammar or phonology (for example, force-fitting morphological paradigms into the traditional katsuyōkei framework, or transcribing vocabulary in kana).

    These two problems (oppression, and description bias) are particularly problematic when the minority languages are facing extinction, as is the case for the entire Japonic family. We want the best descriptions we can get, while there's still time; and we want to give minority speakers all the support we can.

  27. leoboiko said,

    February 22, 2016 @ 5:38 pm


    MSM with Cantonese pronunciation? What is it that I'm reading?

    To complement Eidolon's answer: when some culture uses one language for speaking and another for writing, we call it "diglossia". In diglossia, sometimes the written language may be used as speech, as a formal way of expression, but it's not used in natural conversation.

    One common characteristic of diglossia is that the written language has "parasitic phonology". By this we mean that the pronunciation of the written language depends on the spoken one. In Europe, for example, it used to be that Latin was the written language of choice, even in countries where people didn't speak Latin. But when an Englishman read a Latin text, he'd read it with different pronunciation than a Spaniard or a Russian. Think of the way English speakers today pronounce "Beijing", "Paris", "Mexico" or "Homer".

  28. liuyao said,

    February 22, 2016 @ 6:39 pm

    Thank you all, good discussion.

    I did not fully understand the word "to subsume", and after looking it up I feel you are playing word game here. Alex Lo, who presumably speaks Cantonese and knows the difference between colloquial speech (to throw in another term) and the formal writing, could simply mean that the written Chinese can be read intelligibly by both Mandarin and Cantonese speakers.

  29. Victor Mair said,

    February 22, 2016 @ 8:21 pm

    I don't play word games. That is an unjust accusation.

    If you did not understand what I wrote, you should not have responded repeatedly at such length and with such conviction.

    What I am hoping is that someday Chinese intellectuals will start to take spoken Cantonese (and the other topolects) seriously. After all, the official regulations in Hong Kong stipulate a policy of liǎng wén sān yǔ 兩文三語 ("biliterate and trilingual") — and here I'm using the Chinese terms — which refer to Chinese characters and the Roman alphabet (liǎng wén 兩文) and Cantonese, Mandarin, and English (sān yǔ 三語). Wén 文 refers to script; writing; yǔ 語 — in contrast to wén 文 — refers to spoken language.

  30. liuyao said,

    February 22, 2016 @ 9:04 pm

    Conviction: deeply held belief. I don't think I have or showed any beliefs. I was accepting most of the points raised here.

    [VHM (2/23/16): "conviction" has many nuances. See here, here, here, here, etc.]

    Apology for my strong words. Wasn't my intention. I'll stop short.

  31. Victor Mair said,

    February 23, 2016 @ 12:47 am

    From Robert S. Bauer:

    [As Bob Bauer has been showing in diverse ways for decades, it is possible to write Cantonese. Bob will give a lecture on written Cantonese at a symposium to be held in Japan late next month. The same lecture will be given at Osaka and Tokyo Universities. Here is the abstract.]



    Robert S. BAUER 包睿舜, Honorary Professor
    Department of Linguistics, University of Hong Kong

    Cantonese is Hong Kong’s predominant speech variety: today 90% of Hong Kong’s ethnic Chinese population of about 6.5 million speak Cantonese as their usual, daily language; if we include those people who speak it as another variety, then the percentage rises to 96% of Hong Kong’s total population of just over 7 million. Based on its widespread use by government officials, the broadcast media, and ordinary Hongkongers, Cantonese can be considered Hong Kong’s de facto official spoken language. In 2015 the world-wide total population of Cantonese speakers, including those speaking so-called dialects of Jyut (or Yue, the name of the major Chinese dialect family to which Cantonese belongs), is 62.2 million, according to SIL’s Ethnologue.

    In its Hong Kong environment the Cantonese language has evolved into a dynamic and independent Chinese variety with its own distinctively-defining features. At the outset one thing should be made crystal clear: the Cantonese language is not simply the standard Chinese characters as pronounced in Cantonese. Among all the Chinese varieties spoken in China and across Southeast Asia, Hong Kong Cantonese has uniquely distinguished itself by having developed its own written form that is widely used across a vast range of domains throughout the speech community.

    The phrase《我手寫我口》(ngo5 sau2 se2 ngo5 hau2) ‘my hand writes my mouth’ (which dates from the late Qing dynasty), i.e. I write the way I speak expresses precisely what Cantonese speakers are doing – writing down verbatim with Chinese characters and letters of the English alphabet the vocabulary and grammar of their Cantonese speech. Written Cantonese has never undergone formal standardization; nonetheless, it has accumulated over time relatively consistent conventions which writers adhere to in producing their texts so that they are intelligible to Cantonese-speaking readers. Although these conventions are not explicitly taught in Hong Kong schools, schoolchildren still pick them up informally and so learn to read and write Cantonese through their exposure to its pervasive use. On the other hand, written Cantonese can be almost unintelligible to Mandarin speakers from Beijing or Taibei.

    Why do Cantonese speakers write in Cantonese? Cheung and Bauer (2002:4) have answered this question as follows “. . . writing in Cantonese is perceived by writers and readers as conveying the writer’s message with a greater degree of informality, directness, intimacy, friendliness, casualness, freedom, modernity, and authenticity than writing it in standard Chinese, which is the formal language the Hong Kong Cantonese speaker learns to read and write in school, but its spoken counterpart s/he does not ordinarily use when speaking with coworkers, friends, and family members.”

    Close analysis of Cantonese texts reveals five processes operate in written Cantonese: viz., traditional usage of the Chinese characters, but also their phoneticization, indigenization, and semanticization; along with alphabetization as the result of intimate contact with English. In addition, 10 basic principles have been identified as underlying written Cantonese, and these help us better understand how the five processes operate. Finally, there are two main problems of variation in how lexical items are written in Cantonese which need to be resolved in order to advance its standardization.


  32. Bob Ladd said,

    February 23, 2016 @ 9:30 am

    @leoboiko: I agree that topolect is a good neologism, and I certainly agree that dialect is a loaded term in all kinds of ways, and that lots of governments do nasty or just petty things to enforce various kinds of linguistic uniformity (many mid-20th century Western European school systems certainly did, for example). But I don't think that means that intelligibility alone is a good criterion for deciding whether two related varieties are different languages or not. Perhaps most importantly, as you said yourself, intelligibility is a matter of degree (as any North American who's been to Glasgow or Bombay can attest). It's also (as you also said) subjective, and often asymmetrical (it's probably easier for Portuguese speakers to understand Spanish than for Spanish speakers to understand Portuguese). So if you base your definition on intelligibility, you still won't have a basis for resolving borderline cases.
    Moreover, identity and politics are real factors in the way real language users think about real situations, and there's no reason to pretend they're not. Among other things, it's important not to think of these things only as a matter of oppression of weak groups by powerful groups (or states, or standard languages). There are lots of cases of closely related but never-standardised languages/dialects where speakers insist they can't understand other speakers who come from somewhere else and/or have a different group name even though the objective linguistic differences are small. There are also, as I said in my earlier comment, lots of examples of the converse case, where speakers of very different varieties recognise other forms as being their own language even though they find them difficult to understand. All of these are facts on the ground, and they all need to be taken into account in imposing categorical distinctions ("different languages", "dialects of same language", etc.) on something that inherently involves continua.

  33. Jenz said,

    February 23, 2016 @ 9:55 am

    I think Tatlow's description of Cantonese as a dialect may well be her own rather than a NYT editorial stance. Most educated people (especially those who go through a British educational system) are both institutionally and socially taught to pick up on what constitutes 'good grammar' and proper English, despite there being a splendid variety of prominent English accents that compete with standard Southern English. It leads to an analogy of 'Institutional Language' (Southern English or Putonghua) = standard language = 'real language' vs. colloquial language (most English dialects or Chinese languages) = non-standard Language = dialect, despite the fact that the differences between RP and the Yorkshire dialect are worlds apart from the differences between Putonghua and Cantonese.

    One finds similar things in other cases as well. I'm a (very, very poor) legacy speaker of Turkish. Whenever I meet someone from Azerbaijan (mutually intelligible with Turkish but with its own orthography, literary history, institutional status etc.), I try to make some small talk in Turkish. Most Azeris becomes really shy, citing that they 'don't speak the Istanbul/Ankara dialect,' only the more low-status Azeri. This is despite that fact that my Turkish pronunciation and grammar is very obviously at the level of a slow-witted beginner. This is in pretty stark contrast to Scandinavians, who are more than willing to engage each other in conversation while using their own respective languages (or dialects, however you wish to view it).

  34. Victor Mair said,

    February 23, 2016 @ 4:13 pm

    From Brian Spooner:

    From my own experience I would say that the main problem between Turkish and Azeri is vowel harmony. I could never work out how to get the harmony right in Azeri. But we really need a Turkic specialist, someone like Walter Feldman (rememeber him?) to explain to us why there is such a high degree of mutual intelligibility between the various versions of Turkish/Turkic, from Sarajevo to Turfan.

  35. Jeff W said,

    February 23, 2016 @ 6:35 pm

    From Bob Bauer’s abstract:

    Hong Kong Cantonese has…its own written form that is widely used across a vast range of domains throughout the speech community.

    For someone like me, who really doesn’t know, what does that mean in the real world? Is written Cantonese the “default” for most of what Cantonese-reading people in Hong Kong will encounter—such as in an advertisement, book, magazine or newspaper? Or is it used mainly in places where one wants to convey Cantonese as spoken—for example, dialogue in a comic book, a TV script, or the transcript of a court proceeding or a deposition? Or something else (e.g., text messaging)?

    And, if there is, as seems to be the case, some formal form that is taught in schools and this written form that is closer to the spoken form, what do people call one or the other, if anything? If the implicit point of Bauer’s talk is that this form of written Cantonese is largely unrecognized by people using it (I’m not sure that that is the point), how do you even ask them about it in a way that ensures reliability?

  36. Victor Mair said,

    February 23, 2016 @ 10:04 pm

    From Peter B. Golden:

    The comments regarding Azeri are on target, but it depends on the individuals.

    “Dialect” has become firmly embedded among Chinese abroad when referring to the different topolects (a term I like, but I do wonder if it will catch on). My wife, as you know, is a native of Shanghai (Shanghainese is her native tongue, with some Wuxi touches from her mother), came to Hong Kong as a grade-schooler in 1949 and was unable to understand one word of Cantonese. In school, the principal’s daughter who knew some Mandarin, sat next to her in class and translated from Cantonese until she was able to get the swing of things. She and her younger siblings all acquired fluency in Cantonese, her older siblings speak it with an accent. Her mother spoke it with a really heavy accent and basically translated Shanghainese usages into a kind of Cantonese. A number of her Cantonese school friends with whom we are still close, occasionally kid my wife for her accent. She, apparently, still has some Shanghainese traces – she arrived in Hong Kong at the cusp of the age at which one can still acquire native, accent-free fluency.

    Yes, “dialect” is a relic from an earlier age and most certainly does not describe the situation. Taishan (Toisan) is a dialect of Cantonese. Cantonese is not a dialect of Mandarin, but a self-standing language. The same is most certainly true of the Wu grouping and its dialects (e.g. Shanghainese, Ningbo, Suzhou et al.) I am not a Sinologist, but simply have been observing these linguistic phenomena in my wife’s family and among friends (these included Ningbo and Suzhou speakers) for over half a century.

    I have been interested in the question of dialect vs. language in China since 1961, when Sylvia and I met and I first encountered the phenomenon. I was interested in the way the “dialects” were continually intersecting in her generation (not in the older ones). The younger generation tends to favor one “topolect” over the others. Thus, a sister-in-law (my wife’s younger brother’s wife) of Ningbo origins, born in Shanghai but came to Hong Kong as a pre-schooler, speaks Shanghainese with a strong Ningbo touch, but prefers to speak Cantonese, the language she uses with my brother-in-law and children at home, and customarily uses with Sylvia. A group of nieces and nephews raised in Taipei prefer Mandarin. They understand Shanghainese, which their parents customarily speak with each other, but Mandarin has become the lingua franca in the household. They, of course, now have children of their own, to whom they speak in Mandarin (Taipei variant)…but are now living in Shanghai. It gets complicated.

    I am personally quite familiar with language-mixing among closely related languages. I spoke Russian (or rather a regional variant) with my grandparents. It is actually, a mix of Russian and Belarusian. My grandparents were from rural Belarus’ – from farming and rural artisan families. My grandfather (youngest of ten children) was orphaned at six and never had any schooling. I never learned how he acquired literacy, but he could read (albeit always with lips moving and softly saying the words aloud). He worked (this was in Tsarist times) as a rafter (floating timber down the northern Dnieper), saw-mill worker, rope-maker (an ancestral craft) and with crews painting railroad stations in Belarus’ and more in Ukraine (hence this long preface). As a consequence, his language was filled with Ukrainianisms (my grandmother’s family used to call him “khokhol” – a slightly pejorative term for Ukrainian). In one sentence, he could shift in and out of Russian, Belarusian and Ukrainian. It rubbed off on me and when I speak – if I am not careful – I not only replace the hard “g” of Russian with an “h” ( as in Ukrainian and Belarusian), but throw in regional terms. This Mischsprache they term “trasianka” in Belarusian and “surzhyk” in Ukrainian. As I later learned, this is common among people from the areas where Russian, Belarusian and Ukrainian intersect – although less so with those who are educated. Southern Russian dialects also have “h” instead of “g.” The three Eastern Slavic languages blend into each other in the border zones. The Chinese “topolects” are much more distinct, the differences being greater than even those that separate Russian from Sorbian (a Slavic language still spoken in eastern Germany, which does some interesting things phonologically).

  37. Victor Mair said,

    February 25, 2016 @ 10:01 pm

    From Rachel Kronick:

    I encountered a little of this tension when I was doing medical interpreting full-time. Clinics and even interpreting companies would occasionally not understand what 'Chinese' is, or the difference between Cantonese and Mandarin. That meant I sometimes got called to do interpreting for someone who doesn't speak any Mandarin, and I expect that there were sometimes Cantonese interpreters called for people who speak only Mandarin. When medical care is needed, understanding that Cantonese and Mandarin are different languages becomes not just a matter of linguistic pride, but can even be a matter of life and death.

  38. Victor Mair said,

    February 25, 2016 @ 10:28 pm

    From Robert S. Bauer, reporting from Hong Kong:

    The politicization of HK's languages seems to be all around us.

    Yesterday I was at the HK Univ. of Science and Technology where I attended a lecture on "Kongish" (poster about the lecture is attached for your information).

    As I was leaving, I passed by the students' 大字報牆 "Big Character Poster Wall". One of the big posters was entitled in very large characters "反對科大只推普教中 拒絕矮化廣東話" [which I would translate as "Oppose UST only teaching Chinese through the medium of Putonghua, Reject the diminutization of Cantonese"].

    I didn't have a chance to read the full text of the poster, but a friend who teaches at HKUST said there may be a softcopy of it and will send it to me.

  39. Victor Mair said,

    February 25, 2016 @ 10:45 pm

    Bob Bauer also sent along the pdf of a poster for a public lecture titled: "Kongish Daily and Hong Kong English: A Preliminary Study of an Emerging Variety of English" to be delivered on 25 Feb 2016 at the Center for Language Education, The Hong Kong University of Science and Technology.

    I would have added this to "Kongish" (8/6/15), but comments are closed there.

    If anybody wants to see the pdf, I'll be happy to send it to them.

RSS feed for comments on this post