Trainspotting-like Voices in Chinese
« previous post | next post »
I'm trying to imagine a novel such as Irvine Welsh's Trainspotting being written in Chinese. Trainspotting consists of a number of voices, all of them Scottish-accented in one way or another. It's difficult enough to write in unalloyed Pekingese, for example, much less several varieties of Pekingese or of other Sinitic topolects.
Lǎo Shě 老舍 (real name Shū Qìngchūn 舒慶春, a Manchu of the Sumuru clan; February 3, 1899–August 24, 1966), who was renowned for his novels that contain a conspicuous amount of Pekingese terms, used to complain that it was impossible for him to write many of his favorite Pekingese expressions in Chinese characters. The phonetic flexibility of alphabetic scripts, the ability to write down any sounds that are expressed in the speech of a particular language, is conspicuously absent in writing with characters, which is limited to a fixed (and generally rather limited) number of syllables. We have recently encountered these obstacles in a couple of posts on Language Log: "Surprising Transformations of a Beijing Street Name" and "Russian Loans in Northeast and Northwest Mandarin: The Power of Script to Influence Pronunciation."
In Mandarin, there are a little over 400 different syllables, without regard for tone. When the tones are added in, there are less than 1,600 different syllables, since not all syllable types have all four tones. Cantonese has considerably more different syllables than Mandarin (750 distinctive combinations of initials and rimes, without considering the tones). But this is still far fewer than the 8,000-13,000 or so syllables of English (depending upon how they are counted; theoretically, there might be as many as 80,000 possible syllable types in English, but most of these don't occur in actual speech). When ad hoc spellings for foreign words, onomatopoeia, humorous and dialectal effect, and so forth are added in, the phonetic variations available to an author writing in English seem almost limitless, although phonotactic constraints do serve to keep possible syllable types within bounds. (N.B.: We had a lot of very good discussion about the number of syllables in English around the Language Log water cooler yesterday. I hope that my colleagues will contribute their gems of wisdom on this subject to the comments section of this post.)
In Trainspotting, each character employs a unique mode of narration. In particular, the chapters narrated by Renton are characterized by Scots dialog spelled phonetically and those narrated by Davie are presented in Scottish English, while still other chapters are written in Standard English.
Writing full-blown Cantonese of any sort is an enormous challenge. This is a theme we've touched upon many times here at Language Log, most recently in "Cantonese Blackberry Ad."
Writing a novel like Trainspotting in Cantonese would require the author to come up with devices for distinguishing among Hong Kong patois, Mandarin, Mandarin mixed with Cantonese, with perhaps some Hoishan / Taishan speech thrown in, and so forth. I just don't see how it would work, certainly not in a way that would bring along more than a mere handful of readers who would be able to comprehend what the author was intending.
[A tip of the hat to Brendan O'Kane, Mark Liberman, Adam Albright, Lila Gleitman, Ben Zimmer, Bob Bauer, and Don Snow]
Pflaumbaum said,
March 12, 2011 @ 8:43 am
It's always struck me that there's a paradox involved in Welsh's (and others') transliteration of working-class Scottish accents.
It's presumably supposed to give those generally marginalised speakers a voice. But surely it simultaneously does the opposite – by assuming that the reader is English (or at least posh Scots), and therefore that the book is primarily aimed at those historically dominant demographics?
What I mean is, if a working-class Glaswegian normally pronounces football, say, as ['fʉʔbɔʔ], then doesn't writing it fitba imply that RP ['fʊtbɔːl] is how your idealised reader would otherwise pronounce it?
I don't know anything about Chinese so I'm not sure whether this point applies, especially since we may be talking about entirely different languages there, not just accents.
毛根 said,
March 12, 2011 @ 8:53 am
Many Cantonese speakers I know (none of them linguists but all generally highly educated in the sciences) consider written Chinese to be a "different language" from Cantonese, and perhaps different from Mandarin as well, in the same sense that Mandarin and Cantonese are "different languages". I'd be interested in V. Mair's view on this opinion, as this would basically preclude such transcription of accents (wouldn't it?). I can't imagine Scots accent transcribed into German or French being very recognizable as such. Is this analogy even relevant in this case?
In any case a Chinese "睇火車" would surely make some use of the Latin alphabet as every monolingual(ish) Cantonese HK'er I've ever met makes use of them regularly.
Jenno said,
March 12, 2011 @ 9:12 am
This addresses somethig I hab wodered aboud: is id possible to wride Chidese-with-a-head-code?
Pflaumbaum said,
March 12, 2011 @ 9:15 am
@ 毛根 –
I don't see why a Scots accent couldn't be transliterated into French, any less than an RP English accent can. Here's a stab at it:
Aavez-vu des frréres? ('Avez-vous des frères'?)
As opposed to, say, RP, where you might try to evoke diphthongs for /e/ and a non-rhotic frères:
Aveiz-vous deis frès?
These are probably poor efforts, but I don't see why one would be possible in principle and the other not.
Dan Achibald said,
March 12, 2011 @ 10:07 am
What I mean is, if a working-class Glaswegian normally pronounces football, say, as ['fʉʔbɔʔ], then doesn't writing it fitba imply that RP ['fʊtbɔːl] is how your idealised reader would otherwise pronounce it?
I think what the spelling means to suggest is that Scots is a language separate from English, and not merely a way to pronounce English. There is no standard Scots orthography (though not for lack of trying), but that hasn't stopped authors from Burns to Kelman (and of course Welsh) from doing their best to give it a literary voice.
And then, of course, there's also the question of how to distinguish different voices within a text, which is I think the point of this post. Yes, authors will sometimes sink into stereotypical eye dialect, but I can't think of a better way to do it. "The following section should be read in Glasgow Scots." Maybe with an audio appendix for folks unfamiliar with the dialect? No, I think changing spelling is the best we've got.
Dan Lufkin said,
March 12, 2011 @ 10:18 am
Now is the time to mention the publisher James H. Heineman's 1991 project to publish P.G. Wodehouse's immortal The Great Sermon Handicap in many languages.
I have only the first four volumes covering English, Esperanto, Pidgin, Kréol, Papiamentu, Finnish, Magyar, Basque, Kalderash Romani, Welsh, Breton, Scots & Irish Gaelic, Latin, French, Spanish, Italian, Portuguese, Romanian, Catalan, Rhaeto-romansch, Middle English, Dutch, Flemish, Afrikaans, Friesian, Plattdeutsch, Lëtzebürgisch, Yiddish, Schweitzerdeutsch, Swedish, Danish, Nynorsk, Bokmål, Faroese, Old Norse, Icelandic and phonetic English. There are another four or five volumes in Asiatic and Slavic languages.
Heineman contributed an erudite foreword on the challenge of translating Wodehouse's unique voice into, say, Swahili.
These books are not easy to find, but they're worth while looking for. The quality of the individual translations is generally excellent. This is a publishing tour de force the likes of which we'll not see again.
Karen Kay said,
March 12, 2011 @ 12:10 pm
When I was living in Japan in the 1970's, I read Agatha Christie's "Murder on the Orient Express" in Japanese. This is about a murder committed by a dozen people, and so throughout the book, which is extremely dialogue-heavy. You had to figure out who was speaking not so much by what they said as how they spoke.
Jeff DeMarco said,
March 12, 2011 @ 12:28 pm
The reference to Trainspotting made me think of W.S. Gilbert's The Hooligan, in which he attempts to transcribe turn of the 20th century Cockney speech patterns. You can view it here, along with a transcription into modern English beginning at p. 12.
Bill Ray said,
March 12, 2011 @ 2:58 pm
You can find another great example of dialectical differentiation is Huckleberry Finn, in which Mark Twain takes great care to distinguish between the different dialects of Missouri and western Illinois, as well as the various African-American dialects. Hard to imagine that happening with a character-based writing system.
Not a naive speaker said,
March 12, 2011 @ 3:25 pm
I thought of Feersum Endjinn by Iain M. Banks (another Scot).
Maybe the solution to this problem is not writing but speaking the different varieties. Under Milkwood for example.
Michael Rank said,
March 12, 2011 @ 4:55 pm
For what it's worth, Trainspotting is translated as 猜火车 ("train guessing", which has nothing to do with the ancient hobby of trainspotting).
http://www.flickr.com/photos/ibisbill/5022671814/in/set-72157624902265599/
John Wells said,
March 13, 2011 @ 4:01 am
If we allow 23 initial consonants, 20 vowels, and 23 final consonants (disallowing initial ŋ and final h, but allowing ʒ initially and finally), plus 23 initials x 13 vowels/diphthongs that can be found in open syllables, we get 10,879 possibilities for a variety of English like my own.
Leonardo Boiko said,
March 13, 2011 @ 5:16 am
What about romanization (or other phonological scripts) for Chinese languages? I suppose most of them must have something like pīnyīn by now, right? Would they be able to handle local varieties for added color? Or would the use of them in literature be too audacious?
Bob Violence said,
March 13, 2011 @ 6:07 am
It'd be too audacious, if for no other reason than non-Mandarin romanization schemes are usually unknown even by speakers of those languages/dialects. And plenty of them still don't have a standardized romanization — Wu (the second most common Chinese language family, behind Mandarin) is still usually romanized into the IPA because nothing else has wide acceptance.
Victor Mair said,
March 13, 2011 @ 7:52 am
@Leonard Boiko and Bob Violence
It's not too audacious.
Nearly everyone in China under the age of 40 or 50 is familiar with Pinyin if they have any education at all, since everyone who goes to school begins to learn to read and write using Pinyin. Goodness, I have countless Chinese friends in their 60s, 70s, and even 80s (as well as one who is 105!!!) who are quite comfortable reading Pinyin e-mails from me and writing e-mails to me in Pinyin when circumstances call for it. Furthermore, the vast majority of people in China use Pinyin to input Chinese characters and to do texting. And, as I have shown countless times in my LL posts and elsewhere, romanization is often used directly in transcribing bits of speech mixed in with characters (I could — and should — give many more instances of the use of Pinyin in daily life).
Just last night, a Chinese archeologist friend of mine whom I'm bringing to the United States for a research visit caught an error on the electronic ticket I purchased for him (my travel agent had registered his wife's given name as Liuli instead of Luli). Mind you, this is a person who is in his mid-70s and can neither speak English nor understand spoken English, and who can only with difficulty read a few words of written English.
Then again, nearly every educated person in China under the age of 30 or 40 (and many people older than that) nowadays is conversant with English. Thus, knowing romanization in two different contexts — for Mandarin and for English — makes it all the easier for individuals to engage in ad hoc romanization of topolectal expressions. And don't forget that romanization is widely and commonly used in the writing of Cantonese, while Church Romanization (Pe̍h-ōe-jī — there's a superb Wikipedia article on this subject) — with roots going back to the 16th century — is still extensively used to read and write Taiwanese.
So, there are excellent prospects for the growing use of romanization to write things that it is inconvenient or impossible to write with characters.
David Moser said,
March 13, 2011 @ 8:16 am
Great topic, and one I've been interested in for some time. See my article "Some Things Chinese Characters Can't Do-be-do-be-do" on Mark Swofford's Pinyin.info site for some other examples:
http://pinyin.info/readings/moser/chinese_characters.html
Bob Violence said,
March 13, 2011 @ 9:06 am
Nearly everyone in China under the age of 40 or 50 is familiar with Pinyin if they have any education at all, since everyone who goes to school begins to learn to read and write using Pinyin.
But how suited is pinyin to non-Mandarin languages? I don't know much about Wu, for instance, but Shanghainese at least makes a three-way distinction between plosives (e.g. p/pʰ/b̥) not found in Mandarin; an adapted pinyin-style orthography (like Sichauanese Pinyin) would surely be workable and I've seen ad hoc pinyin schemes for Shanghainese (e.g. bb/b/p), but as things stand, could a writer produce a text in romanized Shanghainese and count on it being recognizable to a Shanghainese speaker? Even Pe̍h-ōe-jī (which has the benefit of being a fully standardized orthography) isn't known to my Taiwanese friends. (Apparently Tâi-lô is now part of the syllabus in Taiwan, so things will hopefully pick up in the future.) Of course none of these issues are reason to simply give up, but it does strike me that writers using romanization for most of the non-Mandarin languages would be pushing the envelope.
Eric Henry said,
March 13, 2011 @ 9:46 am
One question raised here is the potential multivocality of Chinese, and I wonder if case is not somewhat overstated. Certainly Mandarin has less syllabic diversity than English, but in my experience, Chinese readers extrapolate voice and accent from smaller semiotic cues. Use of vernacular phrases or characters (啥 for 什么 or 俺 for 我) index various regional identities and accents, despite the lack of explicit phonetic markings.
Miyako Inoue's work on translation into Japanese shows a similar process at work in, for instance, how the voices of black slaves are rendered into Japanese, using ungrammatical copula endings to mirror their idiomatic English.
Inoue, Miyako. 2003. Speech without a speaking body: “Japanese women's language” in translation. Language & Communication 23:315-30.
Sophie Wei said,
March 13, 2011 @ 10:10 am
Romanization is indeed very useful for transcribing Taiwanese, especially when Taiwanese does not have scripts. For the younger generation, they tend to borrow Chinese characters or another system of alphabetic letters, ㄅㄆㄇㄈ, to transcribe the Taiwanese words or phrases in their emails or for Internet communication. The reason is that they are more familiar with the spelling system of alphabetic letters, ㄅㄆㄇㄈ, since they start to learn it from the first grade in the elementary school while Enligsh letters are learned much later. Another reason is because Taiwanese people would not like to be regarded as a part of China. Therefore, whether to use Hanyu Pinyin 漢語拼音 in Taiwan has been a debate for several years. (Previously several different systems of romanization has been adopted, including Tongyong Pinyin 通用拼音.) Since 2009, Hanyu Pinyin 漢語拼音 is adopted as the romanization system in Taiwan. Perhaps it is because President Ma won the Presidential election and took the office.
Interestingly, in order to maintain the Taiwanese identity, some of the counties in Taiwan, whose county magistrates are DPP members, still choose the romanization of Taiwanese pronunciation for the names of streets, not the Mandarin pronunciation, to put on the signs. This causes quite a confusion for foreign visitors, since their guide books might be filled with information of the romanization of Mandarin pronunciation for the street names.
Romanization can be a means to transcribe what is impossible to write with characters, but it can also be maneuvered as a political tool.
Helen Liu said,
March 13, 2011 @ 10:53 am
@ Bob Violence wrote:
"…non-Mandarin romanization schemes are usually unknown even by speakers of those languages/dialects. And plenty of them still don't have a standardized romanization…"
True, there is still a lack of standardized romanization system for dialects such as Taiwanese, but I would disagree that romanization of Taiwanese is "unknown" to speakers of Taiwanese. Quite the contrary, most if not all church-going Taiwanese have at one point learned luo-ma-pin-yin in order to read scriptures in Taiwanese. In my early years in Taiwan, long before I was able to speak English, I learned this romanization system and read scriptures in Taiwanese. And as Victor points out, its roots go back to the 16th century.
A quick search in google yields interesting examples of Taiwanese represented in Chinese characters with some words romanized because they lack equivalents in Chinese characters. Passages like this, which look like Chinese, would be pronounced in Taiwanese.
http://iug.csie.dahan.edu.tw/iug/Ungian/Chokphin/Lunbun/Huhokengcheng/LMphengim.htm
長久以來，台語文書寫一直lóng是真困擾ê問題。歷史上，書寫台語文ê兩種主流文字，分別是漢字kah台語羅馬字，漢字部分，自以早歌仔冊用漢字書寫，˜-koh大量使用借音字，kàu現在追求漢字本字ê學者所提出ê用字mä無法度一致，另外，台語漢字一字多音ê情形mä真嚴重，koh加上罕用漢字鬥一腳，這對台語書寫是不利ê因素。
@ Leonardo Boiko
In Taiwan, the zhuyin fuhao, or bopomofo, system has been in used for many years and is the preferred way for inputting characters on the computer and for texting on cell phones. The speed with which young people can enter Chinese text this way is amazing to watch indeed! Whole textbooks are written entirely in bopomofo in the early grades in elementary schools before kids master their characters. Today, you'll find shop signs containing obscure Chinese characters with romanization in bopomofo to help the public pronounce these words. And in the election last November, one candidate for political office whose last name was a particularly obscure word romanized it for the public in bopomofo.
Randy Alexander said,
March 13, 2011 @ 11:12 am
@John Wells: You've only covered CVC and CV, but there are also blends on both ends to consider. If it weren't midnight here in Xiamen, I'd do some looking up and figuring, but as it is I'll have to leave that for someone else.
I remember reading a paper (with accompanying data) somewhere a few years ago that put the number of syllables in use in English (not the number of possible syllables) at around 11,000. I also have seen a similar number for Korean syllables.
Helen Liu said,
March 13, 2011 @ 11:31 am
@ Bob Violence
(Apparently Tâi-lô is now part of the syllabus in Taiwan, so things will hopefully pick up in the future.)
Not sure I'd be so optimistic. So many young people today don't speak Taiwanese well, or at all. If they don't even speak it, would there be any reason to WRITE it? And with Ma in office…
Languages rise and fall with politicians.
Rashon said,
March 13, 2011 @ 12:38 pm
This is very interesting. It seems that the use of Taiwanese or Catonese in a Mandarin novel would make it a bilingual novel rather than "accented" or "dialect-ed", and that these local languages might have to be given the same treatment as Japanese or English in a Chinese work. I have no idea where to draw the line of distinction between accent, dialect, and "fangyan" in literature use. Recently I've been researching a paper on Taiwanese literature, and one of the novels I'm analyzing supposedly makes liberal use of Japanese and English letters along with Chinese characters. I believe there is also some Taiwanese or Hakka used in the novel, but I have to check the Chinese version to see if the author uses Chinese characters or roman letters to represent the names. Another interesting thing I thought about is how an author may anticipate a certain reading of a text. In certain Taiwanese novels I imagine an author would have to give textual clues to show they want a character read in Mandarin, Taiwanese, or Japanese. Of course they could use katakana or roman characters to make it explicit, but they might not to prove a point.
Pflaumbaum said,
March 13, 2011 @ 4:32 pm
@ Dan Archibald
And then, of course, there's also the question of how to distinguish different voices within a text, which is I think the point of this post. Yes, authors will sometimes sink into stereotypical eye dialect, but I can't think of a better way to do it.
Well one way is to reproduce the rhythms of the dialect or accent faithfully, along with lexical specificity.
That for instance is what Joyce does much of the time, and his dialogue does feel indubitably Dublin Irish (though I'm sure I miss many of the regional and class nuances).
Ben said,
March 13, 2011 @ 6:53 pm
Hi Victor, I have to agree with Bob Violence on your last comment.
His argument was not whether or not Hanyu Pinyin is well known, but how well suited it is to representing other dialects. Pinyin does not cover the voiceless alveolar lateral fricative of 台山話. But even more importatntly, it does not cover the ng intial in Cantonese. It also lacks the full stops -p, -t, -k.
Also there is a matter of tone diacritics – Hanyu Pinyin will have to be augmented to cover a much larger number of tones.
While an English speaker may not have a hard time deducing the best pronunciation of something like "jaap", a native Mandarin speaker with no previous exposure to Cantonese will probably be confused why their very familiar Hanyu Pinyin is suddenly looking so unfamiliar.
Even if this is not a question of transcribing actual Cantonese dialogues, but just Cantonese-accented Mandarin, I think there is still a task not well suited to the existing Hanyu Pinyin. There would need to be a means of representing 走音.
michael farris said,
March 14, 2011 @ 2:35 am
I'm pretty sure that Hanyu Pinyin can express non-Mandarin accents in Mandarin. Some years ago I had to input several pages of written Chinese from a PRC Cantonese speaker. They provided Mandarin pinyin under each character to facilitate this. It mostly went smoothly enough but I ran into real problems without about 10 per cent of the characters when I couldn't match any of the characters available on the screen with the one written on paper.
I consulted a friend (American) who was fluent in Mandarin and he said a lot of the pinyin was off. The Cantonese speaker had transliterated them with a Cantonese accent (IIRC s,x and sh were often involved). He Putonghua'ed the Cantonese accented pinyin for me and the problem was solved.
But forget Cantonese, can a Mandarin author indicate native non-standard Mandarin?
My intuition is that replication of speech has never never been a concern of literate Chinese at all and replicating different varieties of speech in a single text is completely below the radar in the same way alphabet users are indifferent to things that can be done with characters but not an alphabet.
scav said,
March 14, 2011 @ 4:02 am
@ pflaumbaum, re Glasgow pronunciation.
I'd say fitba is pronounced more like ['fɪʔbɔ]. I'm not 100% sure about the [ɪ], but I've definitely never heard fitba with a glottal stop at the end.
Victor Mair said,
March 14, 2011 @ 6:24 am
@Ben
Not at all.
I published a journal of Romanized Mandarin for a decade, and we found that — if you have proper word division and proper orthography to reflect grammar — you don't even need to indicate the four main tones, much less subtle changes such as tone sandhi. People just read it off with the correct tones, sandhi, and so forth.
It's like the way people read off Russian without the stress marks indicated: they automatically make the necessary accentual and phonological changes as they go, or the way readers instinctively add the required vowels in vowelless Hebrew and Arabic. No writing system can provide all of the phonetic details of real speech. In that sense, all writing systems are more or less schematic. But they do have to provide enough clues to get the reader going and to put him / her on the right track. Cantonese / Taiwanese / Pekingese / Shanghainese / Sichuanese written purely in characters has a hard time doing that, but written in Pinyin makes it a much more realistic proposition.
In this regard, what Michael Farris said above is right on target, though I would add in response to his final, insightful paragraph ("My intuition is that replication of speech has never never been a concern of literate Chinese at all and replicating different varieties of speech in a single text is completely below the radar….") that the reason literate Chinese are not interested in replicating different varieties of speech in a single text is that they don't have the tools for doing it. Pinyin now gives them such a tool, and we see from what goes on in chat rooms and elsewhere that they are gradually starting to use it for that purpose.
David Moser said,
March 14, 2011 @ 7:36 am
"No writing system can provide all of the phonetic details of real speech. In that sense, all writing systems are more or less schematic. But they do have to provide enough clues to get the reader going and to put him / her on the right track."
Ai agree kumpleetly wid Vikta on ths kinda point. A nadiv spieker cn undrstan almst eny kiend of orthografy, sistimadik or cmpletly grbled. We r aymaizinglee gd at deedeucing th sowns of ar naydiv lngwidge, eevn with lawts uf incunsistinzies or errrirz or ohmishnz. IF pnyn wr adoptd, th chaikneez peepl wd kwkly ajust 2 th sistim an mayk it a flxible tool fr cmmoonicashin.
Bob Violence said,
March 14, 2011 @ 10:09 am
I remember reading a paper (with accompanying data) somewhere a few years ago that put the number of syllables in use in English (not the number of possible syllables) at around 11,000. I also have seen a similar number for Korean syllables.
11,000 is the approximate number of syllable blocks that can be formed from the letters currently in use, but I seriously doubt a majority of those actually represent distinct syllables in spoken Korean. This writer claims the South Korean spoken standard has a bit under 3,000 syllables, which sounds reasonable, but I can't find another source for it.
Brendan said,
March 14, 2011 @ 10:37 am
@Eric Henry – The problem with 俺 instead of 我, 啥 instead of 什么, etc., is that they're not actually very useful except as crude ways of indicating that someone is speaking something other than Standard Mandarin. 啥, in particular, is used (with varying pronunciations) all over the country, so a hypothetical reader seeing it won't really get much of a sense of an accent. And among people who use 我 instead of 俺, there's still a very wide range of variation in accents — just think about Sichuanese Mandarin.
I've thought about the problem of indicating accent in literature for a while now, and not much really comes to mind. One way of dealing with it would be furigana, or ruby — small characters printed alongside the hanzi to indicate pronunciation. I remember hearing about a Japanese writer who wrote a book about her African-American boyfriend using the standard 黑人 with furigana indicating the pronunciation "brother." This sort of system is rarely (if ever) used in China, though, and I don't think I've seen it in Taiwanese books other than children's texts providing zhuyin readings for unfamiliar characters; obviously, it would get mighty unwieldy if used more than a few times per page. Even if people were willing to use it, I can't see it being effective for anyone who had as keen an ear for accent as Flann O'Brien or James Joyce or Mark Twain — or Wang Shuo or Lao She, for that matter, both of whom wrote in a dialect allegedly interchangeable with Standard Mandarin.
JS said,
March 14, 2011 @ 11:46 am
In thinking about these questions, we should probably separate the largely ad hoc lexical or spelling adjustments that might be made with an eye to added secai (Faulkner: "I ain't so sho") from the systematic, conventionalized transcription of a particular Xlect.
As Bob and Ben point out, responsible representation of a number of Sinitic languages in our hypothetical novel would require exactly the same number of systems of the latter sort. Each such system would, in effect, engage only native or highly competent users of the language so represented.
In the case of dialect, Pflaumbaum's first remark becomes important to us as linguists. It would be easy for me, as a user of Standard Mandarin, to make adjustments to my Pinyin spelling so as to create a "Taiwanese Mandarin" or "Shandong Mandarin" effect (to this end, incidentally, tone marks would play a critical role.) However, such devices would primarily engage my fellow Standard Mandarin users — those who share my (largely naive) impressions of these dialects.
That is to say, I would be representing dialects as in effect defective variants of the Standard (this being of course just the impression that "Standard" speakers tend to have.) It is easy to indiscriminately add "ain't" to suggest Southern or African American Vernacular English, or "-r" to suggest Beijinghua; it is less easy to master the rules governing negative concord or application of the retroflex suffix such that such adjustments will produce anything more than a crude caricature (not to say that such a caricature could not be of literary merit, of course…)
Brendan said,
March 14, 2011 @ 12:34 pm
But while some spellings will come off as eye-dialect ("sho'"), there are others that will come off as representations of a non-standard version of the language. In Irish English, you'll sometimes see "amn't," which is (I think) more or less transparent to speakers of other dialects. (Less so, perhaps, the Irish "amn't I not?")
Matt Anderson said,
March 14, 2011 @ 11:12 pm
Following up on Brendan's comment from 10:37, Lao She obviously had a great ear for speech, and it's often immediately apparent that his voice is a specifically Beijing one. Here's a random example from his Luotuo xiangzi 駱駝祥子:
他買了十個煎包兒，裡邊全是白菜幫子，外邊又"皮"又牙磣。
He bought 10 fried buns stuffed with outer cabbage leaves; the outside was "pi" (leathery) and gritty.
Three things very effectively mark this speech regionally – the use of 'baor' for bun captures a bit of the sound of Beijing speech, the word pí 皮 is a regional term meaning something along the lines of 'something burnt crispy, then exposed to dampness, which then becomes soft & tough' (please correct me if anyone has a better definition), and yáchěn 牙磣 describes food that is unpleasantly gritty. These terms may have some usage beyond Beijing, but the specific combination seems very regionally specific.
Luotuo xiangzi was published in 1934. Compare that passage to this one (also selected pretty randomly) from Langston Hughes' Not Without Laughter, published four years earlier:
'It ain't "Come to Jesus" no mo' a-tall. Ministers dese days an' times don't care nothin' 'bout po' Jesus. 'Stead o' dat it's rally dis an' collection dat, an' de aisle wants a new carpet, an' de pastor needs a 'lectric fan fer his red-hot self.'
and from another character (in a later section of the book):
'Jimboy's all right, but he's just too smart to do this heavy ditch-digging labor, and that's all white folks gives the colored a chance at here in Stanton; so he had to leave.'
As in Dr. Mair's original point about Trainspotting, while Lao She can certainly do a lot with characters, it just doesn't seem possible to capture the kinds of differences in speech Hughes effortlessly captures here.
U said,
March 15, 2011 @ 3:52 am
@David Moser: this does require *some* familiarity with the writing system. I have a friend who speaks Taiwanese but finds it extremely difficult to read Church romanization (he has family who sometimes write in it). As an aside, he continues to maintain that "Taiwanese is an unwritten language" in spite of this evidence.
I wonder if the Greek and Cyrillic alphabets are sufficiently close to Latin for somebody to read English in them without prior knowledge.
stormboy said,
March 16, 2011 @ 8:07 am
@U: 'I wonder if the Greek and Cyrillic alphabets are sufficiently close to Latin for somebody to read English in them without prior knowledge.'
I would think not (if you're talking about prior knowledge of the alphabets).
Jacob said,
March 22, 2011 @ 9:20 am
I'm a bit late to the party, but I find this particularly interesting. Apologies for the cut and paste from my blog, but it's difficult to handle HTML code in LL comment forms:
I noticed that my wife likes to *type* Sichuanese with her friends. What's most interesting is that they use characters expressly for their phonetic value and ignore their semantic value. A character with a phonetic value (in Mandarin) similar to Sichuanese is used, regardless of its semantic value, to express Sichuanese. That is a rather unwieldy sentence, but rather than parse it (or me rewrite it), it's probably easier to look at some examples.
Character Pron. in Mandarin target meaning
切 qie 去
老 lao 了
黑 hei 很
逗是 dzou si 都是
（也 ye 耶)
(三 sa/a 啊 special to CD)
Examples: 要切耍 要去‘耍’ 玩
走老 走了
There's many more examples, and that's not even mentioning the lexical differences, e.g. 杂个，浪个，莫 etc etc.
I find it interesting that my wife's generation (early 30s) seem to think of Chinese characters as intrinsically Mandarin. That is, the characters' REAL pronunciation is standard Mandarin, and Sichuanese has no writing system.
Returning to the original question: Sichuanese speakers (probably below a certain age) would understand what 我切老 means, but would someone from Shandong or Suzhou or anywhere else?