Chao on the relationship of dialects to the standard

Qian Nairong 钱乃荣 is a pretty big name in the world of Shanghainese. He’s compiled a dictionary or two and is fairly respected for his work with Wu. And he has a blog.

The following is from a post he wrote at the end of December giving different linguists views of the relationship between dialecs/languages (“fangyan,” basically) and Standard Mandarin. They’re a couple quotes from YR Chao 赵元任. More are available here.

Note that in some places I’ve taken some liberty with the translation and in other places an inconsistency with the Mandarin is just my Mandarin being crappy.


“Academically speaking, the standard language is a dialect, what we usually call dialects of the standard are also dialects, and so the standard in a sense is a kind of dialect as well.”

That is, dialects are different varieties within a language and so since the standard constitutes one of these varieties it itself is a dialect.



Dialects are usually said to be within a single family, spread over the geography blending one into the next; As for what constitutes a language, this is affected by the political situation of different branches, and the origin of these languages isn’t a factor. For example with the former use of Latin in Rome, which gradually changed up to the present where we now have many branches, with one factor being government, and another being differences in writing. We think of Italy, Spain, Portugal and France as having different languages… …But if we look at these languages the similarities become clear. We have places like China with several dialects like Beijing hua, Wu, Cantonese, Hokkien and so on, but because China has always used the same script to write all these dialects, on the surface it seems that the differences (between the dialects, in comparison to the Latin languages) are somewhat smaller…., In China, the dialects are all branches from the same source, (and) although the divergence is, at times, extreme, we think of them as dialects of the same language.

I read the last 我们 as “people in general” and not being inclusive of Chao himself, but I could be completely wrong. I have no evidence that he was in any way supporting the idea of Cantonese as a language.

At any rate it’s true of most people that they believe the writing system to be uniform across the board. And I’ve met more than my share of people who believe all of these to be dialects of some language called Chinese.

The part with which I’d have to disagree is that in large part Mandarin feels constructed to me. Yes it’s based on natural dialects and to a large degree. But since it was itself not a naturally developed language, it seems harder to make the case for it as a dialect than, say, Castillian Spanish. Thoughts?

You can read more opinions here at Qian Nairong’s site (also in Chinese).

edit: Updated w/ Sima’s improved translation.

21 responses to “Chao on the relationship of dialects to the standard”

  1. Sima says:


    I suspect you’ve strayed a little toward the end. I would suggest something along the lines of:

    …but because China has always used the same script to write all these dialects, on the surface it seems that the differences (between the dialects, in comparison to the Latin languages) are somewhat smaller…., In China, the dialects are all branches from the same source, (and) although the divergence is, at times, extreme, we think of them as dialects of the same language.

    I think that puts a slightly different slant on things. Sorry if I’ve not quite followed your use of ‘fangyan’, but I hope this makes sense. Equally, please feel free to put me straight!

    I would read 我们 just as we might use ‘we’ in such circumstances.

    If I’m right with my translation of “used the same script to write all these dialects”, then it strikes me that he was being somewhat generous, but you and others will be better qualified to comment.

    Now…”constructed”…I’ll mull that one over.

  2. I’ve updated the post with your translation pasted in place of mine for that last part. It’s what I was going for but yours is clearer/better/etc.

    Regarding generosity, that’s part of why I wonder if with 我们 he’s including himself. With all the work he did on dialects, especially his work on his ancestral Changzhou dialect of Wu, if he did at this time think that they qualified as dialects of a language in the same sense as high vs low German or if he was saying “well come on, French and Spanish are dialects, in a sense”. But then so are English and Hindi.

  3. Sima says:

    I’ve not read enough of him to really have that feel for it, but I think it’s quite reasonable for, say, the scientist to step away from his work and describe how ‘we’ think about the world, even when his specialist knowledge gives him a much more complex or detailed understanding. When he goes home at the end of the day, or when he explains things to the man in the street, he might reasonably accept that he’s part of the ‘we’. He’s both linguist and ordinary guy. But I don’t think this is especially important to understanding what he’s talking about.

    I think the big point is that these are often political distinctions more than anything else. And to a large extent, it doesn’t really matter, does it?

    The lovely point is that, from the first section you translated, academically speaking, 方言 is not simply 地方话. You try telling that to the man in the street, whoever he might be.

  4. Kellen says:


  5. Sima says:


    So might we have a quick look at “constructed” before I try to sleep? I think I kind of know what you’re getting at, but isn’t this just objecting to it not really belonging to a place?

    My own British accent is fairly neutral, though someone with a keen ear might be able to have a stab at where I grew up. It’s certainly not an accent that allows me to instantly fit in to a given place or region. I guess it’s much more likely a marker of social class (or whatever the modern term for that is).

    So, apart from regional associations, how do you see 普通话 as ‘constructed’? If you’ll forgive my asking;)

  6. GAC says:

    I often wonder if 文字 doesn’t maybe mean more than “script” in some cases, but extends to a literary register as well.

    As far as the “constructed” comment, it seems to me that all standard languages are partly constructed. Your example, Castillian Spanish*, has an Academy devoted to it and all of the Spanish language (albeit the Royal Spanish Academy is much more liberal and inclusive than the French Academy). And while English has no single authority, the various standards (Recieved Prononciation, General American, etc.) are partly-constructed standards based on various dialects. People in the Midwest don’t speak the General American that news announcers are taught to use, precisely, but the two are close.

    *Note: “Castillian” means different things to different people. I presume you mean the standard of Spanish used in Spain.

  7. Constructed may be a bit strong. I certainly don’t mean constructed in any Zamenhof-esque sense of the word.

    It’s less an accent thing than that the standardisation process wasn’t, to my knowledge, simply grabbing 北方话 and saying “ok this is the official language now”. It’s to my knowledge not a dialect that naturally occurred but rather a mild mélange of dialects combined to create another one.

    An extreme example might be if the main dialect of Spanish were not just Castillian but Castillian mixed with Catalan and Columbian and Cubano.

    Again, that’s an extreme example to clarify the point, and not actually 100% representative of how I see Standard Mandarin.

    So again perhaps “constructed” is too strong. I just mean “not naturally occurring”.

  8. GAC: yes I mean the standard of Spain as opposed to that spoken in Mexico or elsewhere, and also as opposed to Sevillan etc spoken within Spain.

    And 文字 as the literary register could be a good alternate reading, but then we might have to include Japanese and Korean and others who were also actively using 文言 despite their own languages not being Sinitic.

    And while there’s an academy, it’s one built up around a natural dialect, not one built in order to create the standard dialect.

    Of course all of this could be based on me having a faulty view of the history of the development of Mandarin as a standard, so who know. I’m no expert in that time period.

  9. GAC says:

    I’m no expert on it either. What I have heard suggests that it’s not so much from northern dialects but from a spoken dialect that imperial officials used among themselves and gradually got into writing as 白话文. But that could be a fairy tale — I’ve never read very deeply into it.

    One question I wonder about, is there any official authority in China that regulates 普通话 specifically? I believe I read something about the Ministry of Culture or someone discouraging the use English acronyms (and presumably other Latin-alphabet initialisms as well) in Chinese, but I don’t know any specifics on how that works.

    Re: Castillian comment. Sorry, my first major is Spanish, so I kind of go into a different mode when I think about Spanish language. Shouldn’t have really worried about the meaning.

  10. Kong says:

    I think “not natural” is a bit strong as well. All language that is used among common people is “natural”, because culture comes from the people. As soon as people use a certain dialect as their own, it becomes “natural”. It would be something else if Mandarin were ONLY spoken in formal or government situations, which it is not (if it were, then you would be correct that it would not be a dialect).

    I guess to illustrate my point… If a person is of mixed blood, are they not still “natural”? Sure, they could consider themselves part of either ethnicity, or their own new ethnicity, but either way, they are still natural. Sure Mandarin was constructed out of various regional dialects, but it’s its own dialect now, and even has sub-dialects.

  11. Kellen Parker says:

    Mixed blood would be more analogous to a creole.

    If it was formed by decisions of a committee as to how things would be and how pronunciation would be, and not simply “yeah this is what we’ll use”, then it fits my definition of not natural, though perhaps not everyone’s individual definitions.

    But again if it wasn’t, and I’m mistaken about the formation of the current standard, then I take it back.

  12. jdmartinsen says:

    As it so happens, a monthly history magazine has just published a short account of the development of modern standard Mandarin. The narrative leaves a lot of room for elaboration, but it’s a pretty good overview (however, I’m not a specialist so I don’t know if it’s at all accurate).

  13. Kellen Parker says:

    Thanks JD. I’ll take a look now.

  14. Kong says:

    I really don’t see how it matters if the language was “artificially” constructed by a committee. If the government’s cultural agenda succeeds, and common people grow up learning and speaking Mandarin, then how can you tell them “your native tongue is not natural”? Doesn’t a language being someone’s “native tongue” make it “natural” by default? Are simplified characters unnatural (as a taiwanese might argue)? Even though there are historical roots for simplified characters, they are just as “constructed” as Mandarin.

    As far as creoles are concerned, the only difference between them and languages are political. As mandarin is a mix of regional languages/dialects, it could be termed an official “pidgin”, but as soon as it is learned as a first language by people, it becomes a full creole.

    I guess my point is that the right or not of a language to be equal to other languages is not dependent on its origins, but rather on the descendants of people who use it.

  15. Kellen Parker says:

    Kong: I agree. As soon as a generation grows up speaking it, it’s done. I should have been clear I was thinking in terms of the original onset, or some phase of revision (like in the 1950’s perhaps).

    But yeah once kids learn language X as their native language, it doesn’t matter if it’s Mandarin or Esperanto of Klingon.

  16. Daan says:

    I’ve been reading a book called 方言與中國文化 by 周振鶴 and 游汝杰, which defines 方言 in basically the same terms as 趙元任. Put crudely, it argues a 語言 (in this case 漢語) is basically the reconstructed proto-language of any language family, with its offspring being a variety of different 方言.

    Therefore, all Sinitic languages now spoken, including 普通話, are 方言. With this definition of the difference 語言 and 方言 in mind, I’ve found their analysis is pretty good. At least in this book, the difference between 語言 and 方言 is not the same as the difference between the English words “language” and “dialect” as commonly used. Translating it thus, without reference to the definitions given in the introduction of the book, would be likely to cause misunderstandings about the nature of their thesis.

    I do not doubt that it’s easy to find people who “believe all of these to be dialects of some language called Chinese”, but if you apply the definitions given by 趙元任, 周振鶴 and 游汝杰 to the Germanic language family and say all Germanic languages are dialects of Germanic, that would be like pointing out Germanic languages are the offspring of the (unattested) common ancestor Proto-Germanic. And that somehow feels a lot less controversial to me.

  17. Kellen Parker says:

    Makes sense, under those definitions. The problem I have with that specific view is it means 普通话,上海话 etc are all on the same level. So just for the sake of some sense of organisation to it all, where does that leave 吴语,粤语 et aliae?

    For that matter, there are now plenty of dialects of 普通话 that one could argue are separate form the dialects of 北方话 that may have previously existed in the same geography, though that may be difficult or impossible to prove.

    Argument for its own sake aside, I see where the “Everything is 方言” view is coming from.

  18. Zev Handel says:

    There is another way of viewing the status of 普通话 within the family of Chinese languages. That is to treat it as a koine, the approach taken by the scholar W. South Coblin. A koine (named after the first Greek koine of ancient times) is a compromise of several regional dialects that functions as a lingua franca for communication across a wide geographic area.

    A koine is therefore distinguished from a regional dialect, which is the language of a particular place. A koine generally uses a restricted set of vocabulary that is common to the majority of dialects that contribute to it. By definition, it has no native speakers, but it may be fairly close to some dialects. Particular regional pronunciations of it may have higher prestige status. Also, unlike a dialect, it has no fixed form, even if a standard is defined. This is because speakers of various dialects will approximate the koine to varying degrees, under the influence of their own dialects. A koine is therefore something like an ideal summa of the speech of the people who use it. Koines are natural, not artificial.

    The old Ming-Qing Guanhua was a classic koine. (Nanjing pronunciation was its high-prestige accent, but that doesn’t mean it was identical to Nanjing speech — it was not, maintaining certain distinctions that were not present in that dialect.) You can think of standard American English in the same way. American English is not the language of any one place — it is not Chicago English, or Washington DC English, or Los Angeles English, or even Midwestern English. I am a speaker of this koine, but my speech is definitely colored by my native New England dialect — I speak Standard American English with more vowel distinctions that most speakers.

    普通话 is a defined standard, but it nevertheless functions like a natural koine. Speakers of 普通话 vary tremendously in their realization of the koine, coloring it in various ways with lexical, syntactic, and phonological influences of their native regions.

    A small number of highly educated families may speak 普通话 as a native language, but this is really ancillary to the nature of 普通话.

    So I agree with Kellen that 普通话 should not be at the same level as Shanghaihua or Beijinghua or the Beifang Fangyan considered as a whole.

    One might reasonably ask whether there is a Wuyu koine — a sort of compromise of Northern Wu dialects around Shanghai and Suzhou that has arisen naturally as a common form of speech among people from different regions, even as that speech is is colored in different ways in the mouths of different speakers.

  19. GAC says:

    I think I’ll have to agree with you. Even people who have learned 普通话 as a native language or as their only language have influence from a local 方言. AFAIK, only news announcers and perhaps public officials speak it without this influence.

  20. Daan says:

    That’s a very interesting way of looking at it, Zev. I hadn’t thought about the Chinese language situation in those terms before, even though I do know my fair bit about Ancient Greek…goes to show where a little lateral thinking can get you. Thanks!

  21. Funny you bring it up, as I’ve often wondered about a Wu koine, though not in those terms. It’s what I tend I try to use form when I do bother to speak it at all but that’s mostly because I can’t always keep the dialects straight.

    I’d say a somewhat textbook-Shanghainese would be the best bet there. The books agree when the speakers don’t.

Leave a Reply