Pitch in Korean dialects

« previous post | next post »

From Krista Ryu:

Recently on the internet, there was an interesting photo posted that pointed out the unique feature of Southeastern dialect of Korean:  tones (some scholars call it pitch, as it is different from the tones of languages such as Mandarin).

The internet post had the following photo and a question: "is it true that Seoulites (people from Seoul / users of standard Korean) cannot pronounce these distinctly?"


Modern Standard Korean does not preserve the tonal feature of Middle Korean (used between 13th-17th century). Hence all four of the above numbers (or mathematical expressions) are pronounced exactly the same, and a speaker of standard Korean will not be able to distinguish these four different expressions just by hearing them.

However, the Southeastern dialect of Korean actually somewhat preserves the tonal / pitch feature of Middle Korean, and allows the users to be able to distinctly pronounce these four expressions.

More specifically, the number 2 is read "ee (long e)" and the alphabet e is also read "ee" in standard Korean. Hence, all four expressions become ee ui ee seung (ui= postposition to indicate possession like Japanese の, seung = "power of").

However, speakers of Southeastern dialect can distinguish these four by pronouncing the number 2 with a "low" tone, and the alphabet e with a "high" tone.

Another fun example is the phrase:

가가가가?

In standard Korean, the phrase does not mean anything and will just be pronounced "gagagaga."

However, a speaker of Southeastern dialect can pronounce the first "ga" with a "high tone," second "ga" with a low tone, third "ga" with a "high tone," and the last "ga" with a low tone, to actually deliver the meaning "is that (person) that person (he or she) ?" (is that person the person you were talking about?)

I find it fascinating that, because of the relative geographical isolation of the regions using this dialect due to mountain ranges, certain features of the language were preserved over time.

Observations by Ross King:

It is indeed a bit cryptic, but seems to refer back to something that was hot here in Korea online for a while. I think it has to do with the way Koreans pronounce the "e" of, e.g., 'email', vs. the "이" = 2 = 二 "two" in combination with the way they render, e.g., "e squared" or "2 squared."

Because southeastern dialects preserve older Korean pitch accent, "e" and "이" have different pitch accents, on top of which the pattern for "X squared" is "X의 2승" where "의" is the adnominal particle — pronounced 에 usually. (the 승 must be the sinograph for 'squared', but I'm no mathematician).

These short video clips demonstrate the differences:

경상도 사람들의 2와 e 구분법 – 2의 2승
https://www.youtube.com/watch?v=RVj7FbvTjso

[부산남매] 부산사투리 2의 2승, 2의 e승, e의 2승, e의 e승 구분
https://www.youtube.com/watch?v=6M8jbW_jZMg

2의2승을 읽을때 서울과 경상도의 차이ㅋㅋㅋㅋ
https://www.youtube.com/watch?v=83sI0F_cx00

Comment by Bob Ramsey:

There isn't just one southeastern dialect; there are at least two heard in the Gyeongsang area with salient differences. In North Gyeongsang (such as, e.g., in Daegu), the reflex of Middle Korean "上聲" is a long syllable, while in South Gyeongsang, it's a super-low pitch. Don't know how the differences play out with this little game, though.

Oh, and one more bit of pedantry. The pitches heard in Gyeongsang dialects are in actuality significantly different from those recorded in Middle Korean texts.

Once again, we encounter the phenomenon of peripheral and isolated topolects tending to be more conservative and archaic (cf. Wu, Cantonese, and Min with Mandarin and Smith Island, Maryland with Baltimore or Washington DC).



19 Comments

  1. David Marjanović said,

    December 17, 2017 @ 7:52 am

    peripheral and isolated topolects tending to be more conservative and archaic

    Or at least preserving different archaisms than prestigious or widespread relatives, while indulging in their own innovations.

  2. Jichang Lulu said,

    December 17, 2017 @ 8:35 am

    The hanja for 승 seung 'power' is 乘 (cf. Chinese x的y次(乘)方).

    Somewhat related: the Shanghainese merger of 'eight' and the bound morpheme 'hundred' 八/百 (paq in Simmons' romanisation, where -q is a glottal stop). Just like the merger of 'buy' and 'sell', this hasn't prevented commerce from flourishing in Shanghai, despite what some might too simply, naïvely think.

  3. Jon Forrest said,

    December 17, 2017 @ 10:56 am

    This is somewhat off topic, but I'd like to learn about how modern Korean, as spoken in South Korea, is different from what is spoken in North Korea.

  4. Victor Mair said,

    December 17, 2017 @ 11:07 am

    @Jon Forrest

    These two posts should give you an idea of some of the differences between language use in North Korea and South Korea:

    "Is Korean diverging into two languages?" (11/6/14)

    http://languagelog.ldc.upenn.edu/nll/?p=15534

    "'Bad' borrowings in North Korean" (12/3/16)

    http://languagelog.ldc.upenn.edu/nll/?p=29558

  5. Jerry Friedman said,

    December 17, 2017 @ 11:45 am

    David Marjanović: Or at least preserving different archaisms than prestigious or widespread relatives, while indulging in their own innovations.

    Thanks. From my not even amateur knowledge of linguistics, I've suspected for a long time that dialect differences were often as you describe them.

  6. David Morris said,

    December 17, 2017 @ 3:49 pm

    A few years ago I attended a talk by a guest speaker at my wife's church (a Korean congregation in Australia). I understood absolutely nothing of it, but I noticed at one point the speaker's pronunciation changed everyone was laughing. I asked my wife about it afterwards, and she said the speaker was talking about regional differences, possibly in terms of language and possibly in social terms.

    So, the differences can be heard by a basic speaker of Korean as a second language, and are good for a laugh.

  7. Kyungjoon Lee said,

    December 17, 2017 @ 5:14 pm

    Another difference between Seoul and Gyeongsang province is that in Gyeongsang province, "e" starts with a glottal stop whereas "2" starts with a vowel.

  8. R. Fenwick said,

    December 17, 2017 @ 10:37 pm

    @Jichang Lulu:

    Just like the merger of 'buy' and 'sell', this hasn't prevented commerce from flourishing in Shanghai, despite what some might too simply, naïvely think.

    Absolutely. And in all seriousness, the merger of "buy" and "sell" just results in a new bidirectional verb that might be well-translated as "to conduct commerce". How would that get in the way of any business one might wish to conduct I don't know. No more than it would for Turkish, which doesn't have distinct verbs for "buy" and "sell", usually using almak "take" and vermek "give" for those purposes (and resulting in the compound alışveriş "business, trade, commerce", etymologically meaning "give-and-take", which I find quite a pleasing image).

    That in itself reminds me of a case in Ubykh, which in its most usual idiolect distinguishes a striking minimal pair between ʑʷe "to cook (food) in water, to boil" and ʒʷe "to cook (food) in a fire, to roast". However, in the lect of one speaker recorded in the 1960s that seems to represent the last vestiges of a distinct dialect of the language, the palatoalveolar sibilants have been merged into the alveolopalatal sibilants, leaving that lect only with the single verb ʑʷe "to cook (by any method)".

    [You know, I think we may have reached Peak Nerd here. I never thought I would encounter a paragraph containing the verb phrase "reminds me of a case in Ubykh". (Ubykh was a Northwest Caucasian language with 84 distinct consonants. It has been extinct since 1992. Amazing that anyone should even know about its ʑʷ / ʒʷ merger, let alone perceive a relevance to tone in Korean!) —GKP]

  9. Mark F. said,

    December 17, 2017 @ 10:42 pm

    So, in Standard Korean, how would one distinguish those expressions? Expressions involving both 2 and e are everywhere in mathematics, so it is hard for me to believe there is not some workaround.

  10. B.Ma said,

    December 18, 2017 @ 2:08 am

    I've never had cause (nor do I know how to) say any mathematical terms in Cantonese, but the tones that would be used for e and 2 sound very similar to the southeast Korean pitch.

    e would use the 1st (high) tone and 二 is apparently the 6th (low) tone although if I listen to myself I seem to be using the 3rd (which is still low)

  11. Victor Mair said,

    December 18, 2017 @ 9:37 am

    From the Math subreddit in reddit:

    https://www.reddit.com/r/math/comments/7kfo8l/til_that_in_standard_spoken_korean_theres_no_way/

  12. Kyungjoon Lee said,

    December 18, 2017 @ 1:13 pm

    @Mark F.:
    Mostly through context, like "number 2" and "the e constant." Also there's another way to say "power of 2," which is quite literally "self-product." Power of 3 is called "three self-product" and so on.

  13. Chris Button said,

    December 18, 2017 @ 3:58 pm

    @ R. Fenwick

    [blockquote] That in itself reminds me of a case in Ubykh, which in its most usual idiolect distinguishes a striking minimal pair between ʑʷe "to cook (food) in water, to boil" and ʒʷe "to cook (food) in a fire, to roast". However, in the lect of one speaker recorded in the 1960s that seems to represent the last vestiges of a distinct dialect of the language, the palatoalveolar sibilants have been merged into the alveolopalatal sibilants, leaving that lect only with the single verb ʑʷe "to cook (by any method)". [/blockquote]

    Thanks for sharing that! As with the mention of "buy/sell" above (which reminds me of the "give/receive" 受/授 merger in Mandarin), can we assume that historically there was a more salient feature that distinguished the two? I remember some examples from Northern Chin languages where very closely related words were sometimes only marginally distinctive as a result of sound change and reductions in inflectional paradigms.

    Also, are you using e to represent ə here rather than a surface phonetic reflex? I've never had the opportunity to work on any such languages, but every time I hit that same ə/a ablaut in Proto-Indo-European (known as e/o in such parts) or in Old Chinese / Proto-Tibeto-Burman, I think of how much less of a mountain I would have to climb in Northwest Caucasian languages when trying to convince people of its underlying reality.

  14. 번하드 said,

    December 19, 2017 @ 10:37 pm

    @Kyungjoon Lee:

    Ah, I will have to try "e" vs. "2" with friends from Seoul, Gyeongsangbuk-do and Busan.
    Thank you very much for that insight!

    But, from a mathematical point of view, your other post made me feel slightly unhappy.
    self-product | 제곱 | *to the* power of 2, squared
    three self-product | 삼제곱 | *to the* power of 3, cubed

    A "power of 2", to me, would be 1, 2 ,4 ,8, …

    Now I'm feeling unhappy to have nitpicked. Please ignore at my risk.

  15. Jichang Lulu said,

    December 20, 2017 @ 11:29 am

    Besides the points made on the robustness and redundancy of natural languages, that make mergers buy/sell, give/receive or eight/hundred perfectly manageable, it should be noted that the spoken rendering of mathematical languages is typically ambiguous. The homophony of 2 and e is interesting because of the Korean accentual phenomena, but completely unremarkable as introducing ambiguity. For example, the typical ways of reading mathematical expressions in (e.g.) English are case-insensitive and make parentheses silent. This alone can have fatal consequences.

    In a sense, the way mathematical expressions are read aloud is of no mathematical significance. If forms of 'oral mathematics' were to be interpreted as mathematical theories, the results would typically be strikingly different from what's intended in written descriptions, often leading to triviality or inconsistency. It's easy to see that, if the real numbers 2 and e are equal, the theory of the reals as a Dedekind-complete ordered field becomes inconsistent. But case-insensitivity, silent parentheses and other features of spoken mathematics can also cause all manner of mischief.

    Modern mathematics isn't normally communicated orally, and there is no widely known standard for such communication (although devising one would be trivial).

  16. Victor Mair said,

    December 20, 2017 @ 11:37 am

    Jichang Lulu's incisive remarks about "oral mathematics" might well also be applied to "oral Literary Sinitic / Classical Chinese".

  17. Shihchuan said,

    December 23, 2017 @ 5:23 pm

    In Mandarin, the question would be distinguishing between e^1, e^e, 1^1, and 1^e (thankfully, none of those are really that useful in math I guess). We can distinguish between e^1 and e^e, and between 1^1 and 1^e, because ^e would be "yī cìfāng", but ^1 would be "yí cìfāng" (as the number 1 undergoes additional tone sandhi: 2nd tone when followed by fourth tone or fifth/neutral tone, 4th tone when followed by all other tones). However, to distinguish between e^1 and 1^1, or between e^e and 1^e, one has to rely on context or say "數字 1" (shùzì yī, "number one") or "字母 e" (zìmǔ yī, "alphabet e").

  18. Jonathan Smith said,

    December 24, 2017 @ 11:12 pm

    @Shihfang interesting but surely "e" is usually yi4, and "1" simply yi1 in yi1 ci4fang1 一次方 'power of one'? — that sandhi has some syntactic conditions attached rather than being strictly phonological (one time = yi2ci4 一次, etc.)

  19. Jonathan Smith said,

    December 24, 2017 @ 11:12 pm

    @Shihchuan*

RSS feed for comments on this post