Bill Gates speaks Mandarin

« previous post | next post »

Here's Bǐ'ěr·Gàicí 比尔·盖茨 welcoming visitors to his new blog on the Chinese social network WeChat:

This is what Mr. Gates says:

Nǐ hǎo, huānyíng lái wǒ de Wéixìn gōngzhònghào


"Hello, welcome to my WeChat account".

What's amazing about Mr. Gates' performance is that he delivers it with such aplomb.  There's much to like about his delivery.  His tones are not bad, and his overall intonation is quite good. I've discussed the relationship between tone and intonation on LLog before, and am planning another post on the subject in the near future.  This is one of the aspects of Mandarin pronunciation that is most difficult for non-natives to acquire, so it is remarkable that Gates does it so well.

The single item in his pronunciation that grated on my ears the most is that his "wǒ" ("I") sounded like "woe".  But guess what?  I've heard many a student at 3rd and 4th-year level who says it exactly that way.

Another outstanding feature of Gates' pronunciation is that he breaks up the words in the sentence in their proper units and enunciates them with astonishingly decent cadence.

What all of this demonstrates is the value of Romanization in teaching people how to speak Mandarin (or other Sinitic topolects).  My wife, who was an ardent advocate of Romanization, used to write out short passages of text in Hanyu Pinyin (Mandarin) on a piece of paper and then walk up to people (utter strangers) on the street and ask them to read off what they saw.  Of course, there would be much giggling and coaxing, but most people would relent and give it a try.  When they were finished, Li-ching would exclaim triumphantly, "See, you can read Mandarin and be understood, and you don't need Chinese characters to do it!"  She performed this experiment many times in New York, Philadelphia, Seattle, Chicago, and other cities.  Doing so was one of the greatest pleasures of her life.

Mr. Gates is off to a very good start!

For a newspaper account on this story that I inexplicably missed when it came out in mid-February, see:

"When It Comes to Mandarin, Bill Gates Is No Mark Zuckerberg" (WSJ, 2/13/17)

[h.t. Mark Metcalf]


  1. Colin McLarty said,

    May 3, 2017 @ 10:08 pm

    I am a huge fan of pinyin, but I think there is no chance that Bill Gates taught himself this bit of speech from a pinyin text. He had a voice coach, or several.

  2. Victor Mair said,

    May 3, 2017 @ 10:26 pm

    The thrust of the o.p. is that he was reading from a Pinyin text.

  3. Colin McLarty said,

    May 4, 2017 @ 6:44 am

    Yes he is reading from pinyin. But he made video to publish, knowing his Chinese will be compared to Zuckerberg's and will be criticized as much as Zuckerberg's. He rehearsed the shot and got every kind of professional help with it he wanted, including voice coaching on these specific sentences.

  4. Anthony said,

    May 4, 2017 @ 8:54 am

    Don't language learners use IPA, or is it only those destined to be linguists?

  5. Philip Taylor said,

    May 4, 2017 @ 9:38 am

    Anthony asked : "Don't language learners use IPA, or is it only those destined to be linguists?". Sadly not, in my experience; I and two of my Chinese teachers toured Shanghai's better bookshops to try to find a bilingual dictionary with IPA, and none were to be found. This would have been within the last ten years, so there has perhaps been some progress since then. Pinyin is far better than nothing, but (for example) the number of different sounds that can be (and are) represented by Pinyin "r" means that Pinyin is at best a guide; regular exposure to native speakers is (IMHO) the only way to really master the pronunciation.

  6. Fluxor said,

    May 4, 2017 @ 9:38 am

    What's most grating for me was the last three characters 公众号. I agree with Prof Mair that he's probably reading off a pinyin script, because his pronunciation of 公 (pinyin: gōng) sounds too much like how an American would pronounce that percussion instrument, gong. The tones on 众号 are too off the mark as well, making 公众号 (public account) a bit difficult to understand without subtitles. Nevertheless, a decent effort.

  7. Bruce said,

    May 4, 2017 @ 10:28 am

    Hanyu pinyin, although not IPA, once you learn it, is perfectly suited to MSM: it doesn't require a special character set or input method, yet is based on centuries old Chinese analysis of their sound system and is precise to the point of irritating some beginners.

    On the last point, an example

    Consider 徐 xu2 (rising tone) = slow and 手 shou3 = hand
    To an unpracticed ear, the initial (i.e. consonant) of both words is the same, but pinyin insists on the distinction.

    Similarly, j/zh sound enough alike to require serious work.

    Finally, Chinese pronunciation instruction includes instructions of how the make sounds including position of the tongue and other factors, so mere imitation of what you hear is not required.

    This makes Hanyupinyin as exacting as IPA and just as informative — within Mandarin at least.

  8. Colin McLarty said,

    May 4, 2017 @ 10:48 am

    @Bruce I'm with you on this: "Hanyu pinyin, although not IPA, once you learn it, is perfectly suited to MSM." It is a huge linguistic achievement. Bruce is right you cannot just learn single pinyin letters. You must know how r (and some others) sounds in different contexts. But you do not need to learn how it sounds in each word. When you know the letter after it, you know the sound.

    But you do still have to hear correct speech if you want to sound good speaking MSM.
    For one thing instructions about your tongue and lips are hard to follow–since most of us never see our tongue when we speak and rarely see our lips. And, as Victor Mair says, intonation matters along with tone.

  9. Philip Taylor said,

    May 4, 2017 @ 11:28 am

    Colin McLarty asserted "You must know how r (and some others) sounds in different contexts. But you do not need to learn how it sounds in each word. When you know the letter after it, you know the sound". I would respectfully disagree.
    For the same speaker, the "r" in "rén" (人) and the "r" in "rènshi" (认识) can sound completely different, the first not dissimilar to "r" in English "run", the second closer to the "j" of French "jeter". What I and some of my fellow Mandarin students found fascinating was that although the difference was clear to our English/French ears, our native Chinese teachers (all three from SISU) heard the two sounds as the same. It was only when I got to discuss this with the son of the second of the three that I finally found a native Chinese speaker who could not only hear the difference clearly but was able to point out that "rén" (人) and "rèn" (认) are two completely different words and thus even if they differ only in tone as far as the Pinyin goes, there is absolutely no reason why they should sound the same (modulo tone, of course) …

  10. Colin McLarty said,

    May 4, 2017 @ 11:42 am

    @Philip Taylor is entirely right that "For the same speaker, the "r" in "rén" (人) and the "r" in "rènshi" (认识) can sound completely different, the first not dissimilar to "r" in English "run", the second closer to the "j" of French "jeter"." But then too, in Taiwan the "r" in "rén" (人) is also often like the "j" of "jeter," and I expect that happens in some parts of the mainland too. I believe that for the officially mandated (and tested) Putonghua the "r" of "rènshi" (认识) is the same as that of "rén" (人). I could be wrong about that, but am I?

  11. Michael Watts said,

    May 4, 2017 @ 1:20 pm

    Sure, for the same speaker the "r" of 人 and the "r" of 认 may vary between what an English speaker would perceive as /ɹ/ and /ʒ/. This isn't a fact about pinyin; it's a fact about Chinese phonology — it's also true that the same speaker may show the same variation between 人 pronounced today and 人 pronounced tomorrow, much as a cantonese speaker may freely vary between pronouncing 能 with an [n] onset or an [l].

  12. David Marjanović said,

    May 4, 2017 @ 3:56 pm

    Gates delivers his line with great confidence, but basically uses the English sound system, with very little regard for tones, the usual misunderstanding of -ong and a very English misunderstanding of zh (the last two clearly Pinyin-based). I can't see what's impressive about that, except the confidence.

    The sound spelled r in Pinyin is, at the beginning of a syllable in a CCTV accent, intermediate between the retroflex approximant [ɻ] and the retroflex fricative [ʐ]. The former is within the range of English r, the latter is similar to French j (and therefore spelled j in Wade-Giles). I haven't noticed a difference between different words, but there's certainly a difference between accents and probably also between positions in a syllable (beginning or end).

    IPA is very little used in language learning in general. There's a lot of basic linguistics that lots of people would profit from knowing, but it's never taught…

  13. Victor Mair said,

    May 4, 2017 @ 4:00 pm

    Compared to other complete novices, it's very impressive. Many students carry the English sound system through years of learning Mandarin, and many of them never leave it behind.

  14. Anonymous Coward said,

    May 4, 2017 @ 6:39 pm

    probably also between positions in a syllable

    Especially vis-à-vis the position of the syllable in a word. To my ear, for example, anything less fricative than the halfway sound you described is wrong in Anlaut. I would parse the accent as Malaysian, if with Malaysian intonation and vocabulary, but otherwise wrong. In the middle of a prosodic word, even myself can just say a [ɻ].

  15. said,

    May 5, 2017 @ 11:10 am

    To an extent, Chinese sounds (Pinyin) can be used to speak English (see … Gates uses English sounds to speak Chinese… The compatibility of certain words across the two languages is an area deserving of further study!

  16. Drin Ghuen said,

    May 5, 2017 @ 11:52 am

    好 xaw>haew
    欢 hwan>wan
    信 ɕin>ʃin
    公 kong>kɒng
    众 ʈʂong>ʒɒng
    号 xaw>haew

  17. B.Ma said,

    May 5, 2017 @ 4:29 pm

    The initial inability for the native Chinese to differentiate the 'r's in renmin vs renshi reminds me of the puzzle in English which asks readers to count the number of 'f's in a certain sentence, which contains numerous instances of the word 'of'. The English speaker is supposed to neglect the 'f's in 'of' because they are pronounced with a 'v' sound.

    Likewise, native English speakers are often surprised when it is pointed out that 'tree' should really be spelled 'chree'. At least I have never heard a native Eglish speaker pronouncing the 't' and 'r' individually.

RSS feed for comments on this post