Wednesday, January 2, 2008

Perfect pitch — but at the cost of expressiveness?

NB: I’ve been meaning to post about this for ages, so I have to apologize for the staleness of some of the underlying stories I’ll be referring to.

Some time last year, I heard an intriguing piece on NPR (which I can’t find now — maddening!) in which the phenomenon of perfect pitch was correlated with native fluency in tonal languages such as Mandarin Chinese and Vietnamese. Diana Deutsch of the UCSD Pyschology Department is probably the leading researcher in this field, and she’s discovered some very intriguing things. (Here’s a link to one of her most recent, and highly readable, publications on the subject, from the 2006 issue of Acoustics Today.)

Just what sorts of discoveries has she made? How about this? While only about 14% of American English speakers exhibited perfect pitch, fully 60% of Mandarin Chinese speakers did so. That’s pretty staggering, don’t you think? And what’s more, Deutsch recorded Chinese speakers pronouncing the same word on different days, and a different times of day, weeks apart, and the recordings sounded identical (pitch-accurate within a semi-tone — basically, the difference between Middle C and C#). The NPR piece played a number of these samples, and others are available on Deutsch’s website.

No wonder so many Chinese speakers — like my favorite cellist, Yo-Yo Ma (who reportedly does have perfect pitch) — are gifted musicians, right? But after ruminating on this research for a little while, I began to wonder whether this perfect pitch might come at a cost in expressive flexibility.

Think about it. Tonal languages like Chinese convey lexical meaning through precise tones. Why tones? One theory holds that they were necessary for expanding the range of meanings possible in languages that are largely monosyllabic, and often phonemically limited. So, pronounce the same collection of phonemes in each of Mandarin’s four tones, and you get four distinct meanings — most often, completely unrelated to one another. The pitch becomes crucial for understanding. And the more tones you have, the greater the need for pitch-perfect speech. Mandarin Chinese only has four tones, but Vietnamese has six, and Cantonese has nine!

In most Indo-European languages, lexical meaning is unrelated to tone or pitch. Rather, these factors can be used to convey a wide range of secondary connotations instead — irony, humor, emphasis, irritation, accusation, you name it. This is very common in expressive languages such as English and Italian. Consider an example. I just found a more or less random sentence in a recent email (from my wife to one of our friends): “I just wish Jason didn’t have to go to work tomorrow.” Here are a few (by no means all) of the variations possible:

1. I just wish Jason didn’t have to go to work tomorrow.
[Maybe he doesn’t mind, though.]
2. I just wish Jason didn’t have to go to work tomorrow.
[Emphatic.]
3. I just wish Jason didn’t have to go to work tomorrow.
[However I don’t care if you have to.]
4. I just wish Jason didn’t have to go to work tomorrow.
[Alas, he does.]
5. I just wish Jason didn’t have to go to work tomorrow.
[However he is, in fact, obligated.]
6. I just wish Jason didn’t have to go to work tomorrow.
[Rather than work from home, maybe.]
7. I just wish Jason didn’t have to go to work tomorrow.
[Rather than have to go somewhere else.]
8. I just wish Jason didn’t have to go to work tomorrow.
[Today or the day after tomorrow, say, would be better.]

Would all of these shades of meaning be possible in Chinese? I don’t think so! Because of the lexical significance of its tonality, spoken Chinese approaches melody (I tried to capture this in the picture I made, above) — and you can’t alter a melody without producing a completely different sense. Mozart isn’t Debussy. Tonal languages perforce become a sort of monotone. Well, not a literal monotone (perhaps we could call it a “fixed polytone”?), but anyone speaking the same sentences would sound much more “identical” than in English, Italian, or Dutch, for example.

So far, studies have been limited to just a couple of Asian languages. I’d like to see the investigation expanded to consider other tonal languages, such as Yoruba, Cherokee, Hausa, and Punjabi. I’d also like to see the study applied to Indo-European pitch-accent languages, such as Norwegian, Swedish, and Serbo-Croatian. Cherokee might make an especially revealing study, because it’s spoken with tone in Oklahoma but without tone in North Carolina. Unfortunately, the population of speakers is very small (fewer than 25,000, 99% of whom speak it only as a second language).

12 comments:

  1. http://www.wnyc.org/shows/radiolab/episodes/2007/09/25

    ReplyDelete
  2. That's a link to the NPR program

    ReplyDelete
  3. Thanks, Matthew! That must be it! I wonder why I couldn’t find it again through the main NPR website, though. I did find some previous coverage of Dr. Deutsch’s research on NPR, but it was all too old to have been the piece I heard. Thanks again.

    ReplyDelete
  4. Speaking of expressiveness in voice, did you catch this article on Boing Boing?

    ReplyDelete
  5. No, I must have missed that one, but it’s very interesting. It isn’t too surprising, though, is it? I think conventional wisdom would have ventured the guess that we can all tell when a speaker is smiling. Nice to see that confirmed scientifically.

    You know, further to the topic of my post, I was thinking: in English, we indicate a question through pitch as well. How do they do it in tonal languages? :-/

    ReplyDelete
  6. In Mandarin, I believe you indicate a question by the particle "ma" at the end of the sentence. For instance:

    Ni hui xuo pu tung hua = you speak Chinese
    Ni hui xuo pu tung hua ma = Do you speak Chinese?

    ReplyDelete
  7. Oh, and IIRC, there's another way to form questions in Mandarin:

    Ni yao gen wo chih fan = You want to eat with me

    Ni yao bu you gen wo chih fan = Do you want to eat with me? (yao bu yao = want not want)

    ReplyDelete
  8. Ah yes, particles. That’s interesting, and similar to Japanese, where I believe you add か /–ka/ to the end of the statement to form a simple interrogative. I always retained more Japanese than Chinese (but still not that much) from our days of browsing the Teach Yourself’s back in the early 80’s. :)

    I remember you were always better with Chinese anyway, and I guess you learned a good deal more with your recent tutoring gig(s)?

    ReplyDelete
  9. I don't know how you ever got anything out of that Teach Yourself Japanese. I was looking over my copy just a couple weeks ago and found it impenetrable. Just an absolute mess of confusing grammar rules and precious few dialogues to give context. It's really amazing that we learned "as much" as we did back in the day, with that series in tow.

    ReplyDelete
  10. I was looking over my copy just a couple weeks ago and found it impenetrable. Just an absolute mess of confusing grammar rules and precious few dialogues to give context.

    You mean that isn’t what Japanese is really like? ;) Well, at least I learned “koko ni pen-ga arimasu” — which ought to be useful if I ever find myself in a stationery store in Osaka.

    ReplyDelete
  11. Whoever created the music scale with the mandarin characters is pretty impresive! Very creative, very clever, very cool....

    ReplyDelete
  12. Aw shucks, ’tweren’t nothin’ </falseModesty> ... but thanks for noticing! ;)

    ReplyDelete