Welcome to China

« previous post | next post »

In "Doubletalk of the month", Mark Liberman presents a virtuoso display of a woman skillfully mimicking the sounds and intonations of numerous languages.  You can do this kind of imitation with written forms as well.

For example, here is a takeoff on Chinese seasonal couplets:

subway, railway, highway, way way to die

officer, announcer, professor, sir sir to lie

Welcome to China

According to the website where this couplet was posted, it is supposed to have been written by a foreigner, but judging from the way in which it is introduced and commented upon, my guess is that it was actually written by a Chinese person imitating a foreigner (lǎowài 老外 [I will make a separate post on this term]) who is attempting to write an English couplet that conveys the flavor of a Chinese duìlián 對聯 ["matching couplet"]).  Specifically, this imitation couplet takes the form of a Chinese chūnlián 春聯 ("spring couplet"), with the following parts:

shànglián 上聯 (first line of the couplet)

xiàlián 下聯 (second line of the couplet)

héngpī 横批 (horizontal inscription at the top)

Such couplets used to be ubiquitous on the gates / doors of Chinese homes and in more traditional areas still are.

I remember a poignant scene from "Yellow Earth", a 1984 film by Chen Kaige, where poverty stricken peasants on the loess plateau of Shaanxi Province pasted couplets on their doors.  Unfortunately, because they were illiterate and couldn't afford to hire even a low-level scribe, their couplets looked like this:


O                    O

O                    O

O                    O

O                    O

O                    O

O                    O

O                    O

That is to say, they drew empty circles to stand for the characters that they didn't know.

Returning to the English couplet "à la chinoise" at the beginning of this post, this is an instance where mimicry reveals dimly some essential features of the genre being imitated (antithesis, heptasyllabicity / seven words, grammatical parallelism, etc.).

[hat tip Wicky Tse]


  1. Stephan Stiller said,

    March 10, 2014 @ 4:22 pm

    1. This also goes to show that it is much easier to imitate form than content.
    2. In fact, entire genres are (in popular understanding) defined by form only, and this leaves me wondering what the point of such a genre then is.
    3. Consider how the Western understanding of what a haiku is is looser than what it is for Japanese ones. Most conspicuously, there doesn't seem to be a constraint on having exactly 17 "moras" (≈ syllables).
    4. Is it normal that genre developments and adaptations tend to retain form, not content? I don't know; I think that there are many counterexamples in the field of classical music, for example. But then, many genres of (classical and modern/popular) music are defined by rather simple criteria: instrumentation, type of beat, and so on.

  2. Stephan Stiller said,

    March 10, 2014 @ 4:31 pm

    PS: Or maybe there is a bias in that it is much easier to define or describe form than content.

  3. Ralph Hickok said,

    March 10, 2014 @ 6:29 pm

    @Stephan Stiller:
    I'm puzzled by your statement that "entire genres are defined by form only". What do you mean by "genre?" To me, it means something like Gothic romance or historical fiction or mystery, which are surely defined by content rather than form, since all of these and other genres take the novel as their form.

  4. Stephan Stiller said,

    March 10, 2014 @ 8:41 pm

    @ Ralph Hickok
    Good points; looks like that was poor word choice on my part. I was more thinking of what you call "literary form", but then that particular statement of mine looks tautological. And I was in fact also thinking of "mystery novel" as a category, which is certainly defined by content. Talking of which, the 17-mora criterion above is a characteristic of form (not content) which is dropped or much relaxed in Western understanding.

    But it will still be interesting to look to what extent categories (of whichever art form and on whichever level of your preferred ontology and irrespective of what you call them) are defined by form rather than content.

  5. Ran Ari-Gur said,

    March 11, 2014 @ 1:27 am

    @Stephan Stiller: I don't understand your comments about Western haiku. In my experience, they always have exactly 17 syllables. This rule isn't any "looser" than or "relaxed" from the Japanese 17-mora rule, it's just a slightly different rule.

  6. Stephan Stiller said,

    March 11, 2014 @ 2:32 am

    @ Ran Ari-Gur
    Hard to argue without a corpus, and I'm not an expert. Following the Wiki links, the Haiku Society of America says this and this. That is consistent with what I wrote. It is very much possible that my impression comes from the fact that [m]any so-called "haiku" in English are really senryu. If you tell me that haikus produced by literary circles tend to be stricter with this sort of thing, I stand corrected … though it does seem to be the case that what you popularly see labeled as "haiku" is looser. In any case, I'm glad that my views are being challenged. It is also possible that "popular understanding being looser with such things" is a factor at play. And that "haikus" in English are often senryu(?s) is also interesting in and of itself.

  7. Michael Watts said,

    March 11, 2014 @ 3:58 am

    I wouldn't even go as far as to say 17 morae is a different rule than 17 syllables; I'd just say English and Japanese have different phonolgy. When I was instructed in "haiku" in US schools, the only constraint defining them was that they had exactly 17 syllables. Seasonal references were not mentioned; I'm not actually aware of what the formal Japanese rules are.

  8. Stephan Stiller said,

    March 11, 2014 @ 4:28 am

    @ Michael Watts
    Interesting; in Japanese the 3 lines are 5-7-5 moras in length. If haikus are generally taught in the way you describe in US middle or high schools, such "US school haikus" do have fewer constraints (with the only constraint being one about form). By the way I'm with you that an English 17-syllable rule is substantially the same as the Japanese 17-mora rule.

  9. Matt_M said,

    March 11, 2014 @ 10:08 am

    Regarding comparisons between English haikus/senryus composed using a syllabic metre and the moraic Japanese original form, it seems to me that the Japanese metre is much more constraining than the English imitation.

    There are only 100 or so possible Japanese morae, while there are thousands or tens of thousands English syllables. A Japanese mora therefore can encode only about 7 bits of information, while an English syllable can encode perhaps double that (give or take a bit or two). In other words, English haikuists can squeeze many more morphemes into a haiku than a Japanese one can.

  10. Tom Wilson said,

    March 11, 2014 @ 12:25 pm

    On the more general topic of doubletalk-like phenomena, where does cholo writing fit? My entirely casual understanding is that it is a form of graffiti used to mark territory and similar group-boundary functions but is not in fact an encoding of oral language, except perhaps for occasional inclusions of actual words (generally Spanish).

  11. Victor Mair said,

    March 11, 2014 @ 12:41 pm

    Here are some links for cholo writing

    Cholo Writing: Latino Gang Graffiti in Los Angeles (Dokument Press)


  12. Jerry Friedman said,

    March 11, 2014 @ 4:57 pm

    My feeling is that when English-speaking people who aren't serious poets write what they call haiku, they follow the 5-7-5 rule for syllables but don't follow any other rules. That's what I was taught in elementary school.

    When serious English-speaking poets write what they call haiku, they just write poems that are really short and have a similar feeling to (what they perceive in) Japanese haiku, but otherwise with no rules.

  13. cameron said,

    March 11, 2014 @ 6:38 pm

    I wonder if an English poets have dabbled in what we might call "sprung haiku"? I can envision a form that uses a 3-5-3 pattern of stressed syllables, with no restrictions on the number of unstressed syllables. Many other patterns are possible, of course.

    Does anyone know?

  14. Matt said,

    March 11, 2014 @ 11:57 pm

    Does anyone know?

    Yes, a 2-3-2 pattern of stressed syllables, with no restriction on unstressed syllables as you say, is a pretty common approach to English haiku. It doesn't really make sense to simply substitute English syllables (stressed or unstressed) for Japanese morae — obviously there's nothing wrong with a 5-7-5-syllable form in English, but owing to the differences between the languages it doesn't really "feel" the same as a 5-7-5-mora form in Japanese, for those who speak both languages. In particular you can often fit a lot more into a given number of English syllables than the same number of Japanese morae, and in a haiku more is not more.

    There are other formal features of haiku that are difficult to duplicate in English, notably kireji — although given that haiku is a mostly written form some poets are happy with the idea that punctuation in English is "equivalent enough" to what kireji does in Japanese.

  15. Stephan Stiller said,

    March 12, 2014 @ 1:54 am


    In some linguistic analyses, the position of the downstep in Japanese is lazily termed "stress". I think an English 2-3-2 pattern counting stress would correspond to a Japanese 2-3-2 pattern counting the prosodic units of pitch accent (which are inflected words, like one would expect).

    I would think that the reason for why some poets want English haikus that are shorter in syllable count is what user "Matt_M" spelled out. Though that applies to the spoken language. I don't know how densely and creatively kanji are used in Japanese haikus; I would think that one can pack more (unintelligible-if-spoken) information into a 17-mora text relying on kanji than into a 17-syllable text in English, but perhaps someone else can develop that thought.

  16. Matt said,

    March 12, 2014 @ 6:10 am

    Interesting idea! It depends on what you mean by "relying on kanji", really. The link between kanji and pronunciation is so flexible that theoretically you could use the word "kare" ("he") in a haiku, but spell it with kanji specifying that he is an aged traveler from the south of France — but that would be an extremely unusual thing to do, not just for haiku but under any circumstances.

    The vast majority of haiku have always been written in unremarkable orthography, without using kanji to encode anything that couldn't be understood from the text spoken aloud. (Traditionally, when haiku writers have wanted to give more context, what they did was give the haiku a title or embed it in prose, rather than using orthography to game the form.) As a result, on average, they do tend encode less information than you could fit in 17 syllables in English (for the reasons Matt_M describes — and as you astutely observe, when you compare words vs words as opposed to syllables vs morae the numbers do tend to converge).

  17. Stephan Stiller said,

    March 12, 2014 @ 7:49 am

    Reading what you just wrote, I am also thinking that dense kanji usage could be confusing and ambiguous. I mean, it'd indeed be odd to game the 17-mora requirement through esoteric readings (even if we stay with the readings normally associated with any given character) giving rise to ad-hoc Sino-Japanese compounds which might in effect require furigana (for the other readers: phonetic annotations in kana) to verify the 17-mora constraint. Btw how common are furigana in poetry? Is furigana use regarded as clever, clumsy, practical, avant-garde, recommended, recommended against?

  18. Matt said,

    March 13, 2014 @ 9:52 am

    Well, I just flipped through an issue of Gendaishi techo and couldn't find any furigana at all. That doesn't surprise me; it's a literary magazine and the editors probably assume that anyone reading it in the 21st century doesn't need furigana. Other editors, editing for different audiences, might have different policies, but I suppose the general preference would be to avoid it, if only because it clutters the page — unless of course the author is using it for some special effect, but that is quite rare in my experience.

    On the other hand, you see a lot of furigana in editions of older (but still, say, post-Meiji) poetry. Before WWII the orthography was less standardized; authors had more freedom to play around without seeming precious, and furigana facilitated that. At the same time, because the audience, even the audience for "high literature", was not expected to have been as rigorously educated as today's schoolchildren are, even books like Soseki's tended to be published with furigana on every non-numeral kanji, which treatment is generally reserved for children's literature today.

    If you go further back than that you encounter the phenomenon of modern editors supplying kanji in furigana, to clarify words and phrases written entirely in kana that might be unclear or opaque to the modern reader.

RSS feed for comments on this post