Tones in Chinese Songs
I’ve been asked a number of times: if Mandarin Chinese is a tonal language, what happens when you sing in Mandarin? Well, the answer is the melody takes over and the tones are ignored. Pretty simple.
However, it may not quite end there. I recently discovered a paper called “Tone and Melody in Cantonese” which asserts that Cantonese tones are set to music in a somewhat different way:
> For Chinese, modern songs in Mandarin and Cantonese exhibit very different behaviour with respect to the extent to which the melodies affect the lexical tones. In modern Mandarin songs, the melodies dominate, so that the original tones on the lyrics seem to be completely ignored. In Cantonese songs, however, the melodies typically take the lexical tones into consideration and attempt to preserve their pitch contours and relative pitch heights.
Here’s a graphical representation of Cantonese tones, with and without music:
And here’s an example of Mandarin:
I can’t say I’m fully convinced by the pitch contour graphic that the Cantonese songs “take the lexical tones into consideration,” but it’s an interesting argument. This would suggest that studying songs would be more beneficial to acquisition of tones for the student of Cantonese than for the student of Mandarin.
If you’re interested in this kind of thing, Professor Marjorie K. M. Chan has lots of articles available on her website’s Publications page.
Perfect timing! Was just wondering about this myself the other day while listening to some Mandarin songs.
I thought some of the tones in the songs were “incorrect”.
As someone grew up in Guangdong listening to all the Cantonese pops and Mandarin pops, I can confirm that Cantonese songs “take the lexical tones into consideration” heavily. In fact, the affect is so grave that the songs can become ultimately incomprehensible if out of sync with the lexical tones, and it’s really hard to find skilled lyric writers. Cantonese pop songs tend to have very awkward, strange and illogical lyrics, thanks to the new generation of the fast lyric writers who have to trade the quality for speediness.
On the other hand, in Mandarin songs lexical tones are not “completely ignored”, though the lyric writers do not have to consider them consciously. Songs that completely ignore lexical tones sounds a bit strange, and that’s a sign the songs were originally written for another language/dialect, such as the “Happy Birthday” song.
Alex,
Thanks a lot for sharing! This is really interesting, that not only does Cantonese “take the lexical tones into consideration” (heavily), but that listeners also notice it. Would you say you’re atypical in that you notice it, or is this something that most people would notice?
(I need to meet more Cantonese speakers…)
Hello! Cantonese speaker here.
Ironically, my mother who is totally not tone deaf is totally tone deaf when it comes to Mandarin. And even more ironic, she learned Mandarin from listening to mandarin pop songs. Perhaps we’ve just found an answer for her crap Mandarin! 🙂
M
I’m not atypical. In fact this has been a serious problem for Cantonese pop songs as it’s really rare to find Cantonese singers that are capable of composing, writing lyrics and singing. You’ll find in many pop songs inappropriate words or ridiculous words being used simply to satisfy the lexical tone requirements.
Alex, I’ve heard several Mandarin speakers say the same thing, namely that recent pop songs are getting harder to understand because the songwriters don’t manage to convey the tones.
Ah! 90% of my chinese listening is music. Are tones followed in rap lyrics? (Beijing rap)
Usually, I think. Since rap vocals are, more or less, spoken word with an element of (usually 4/4) rhythm, other aspects of speaking are usually preserved. That one mini documentary about Chinese hip hop on Youtube includes an interview with an emcee who mentions having the option of eschewing the tones in order to achieve a particular style of sound, but most of the Chinese hip hop I’ve heard, especially the Beijing Stuff like Yin Tsang and Dragon Tongue, seems to keep the tonal aspect pretty intact.
I just found a recent article about Mandarin that tries to answer this question from the perceptual side, i.e. “how is it that listeners can understand most song lyrics despite the apparent absence of linguistic tones?”.
Basically, they conclude that the syllables that are most prominent in the melody almost always do preserve some indication of their tone, either in the contour or in the relationship to neighboring syllables. But the tone information is incomplete, for instance it mostly just distinguishes “high” tones 1 & 2 from “low” tones 3 & 4. And it doesn’t apply to every single syllable. So it makes the song more understandable as a whole, but it won’t actually tell you the tone of an individual word if you have no guess as to what word it is.
That study was done by looking at the melodies of Mandarin folk songs. It would be interesting to repeat it with pop songs, or in Cantonese.
The article is:
Wee, Lian Hee (2007) Unraveling the Relation between Mandarin Tones and Musical Melody. Journal of Chinese Linguistics, 35 (1). pp. 128-144.
For any linguists out there who might actually want to read it, it’s accessible at: http://eprints.hkbu.edu.hk/166/
SKP,
Wow, that sounds really interesting! Thanks very much for sharing.
My guess is that the fewer tones a language has, the easier it is for it to ignore lexical tones in song lyrics. So I expect Vietnamese and Thai are similar to Cantonese in that they have to follow the lexical tones of the lyrics closely. Any Vietnamese or Thai speakers?
I notice it when Cantonese songs don’t respect tones.
I remember one song for children in both Mandarin and Cantonese versions (don’t remember the name). It was first composed in Mandarin, and sung to a slightly cheesy classical tune. For the Cantonese version, they just used the same characters and sang it in Cantonese—which sounded atrocious and barely comprehensible.
Then again, the “Happy Birthday” song is sung like that in Cantonese: doesn’t respect the tones and some syllables sound ‘off’–but everyone knows the song anyways, so I guess it’s different.
I’m Cantonese. I can tell you that if a Cantonese song don’t take the lexical tones into consideration, it would sound really weird and be considered not professional.
I’ve heard that Taiwanese (闽南语/台语) song has similar requirement.
John, do you have any idea why Cantoese pop songs mostly use Mandarin vocabulary rather than “correct” colloquial Cantonese? (e.g. 的 rather than 嘅 and 不 rather than 唔).
To a learner of spoken Cantonese, it sounds odd to hear “Mandarin” word usage with Cantonese pronunciation, since you would never speak this way on the street.
I asked a friend of mine about this before (I found it odd myself), she replied that it’s because songs are considered similar to poetry, and thus is composed in written chinese and not colloquial chinese.
Like all other have said here, a really good lyricist does take tones into consideration. This is in fact a basic requirement. The so-called new lyricists in Hong Kong are disasters. Mandarin lyrics may be less problematic because it only has four tones. If you study the Cantonese lyrics written by the really good lyricists like 黄霑 or 林夕, they never take the issue lightly. You may also check a song having both Mandarin and Cantonese lyrics, e.g. 约定 by 王菲 or 在晴朗的天空下/真心英雄. The Mandarin lyrics are almost certainly not usable for singing in Cantonese simply because of the tones.
Lyrics in colloquial Cantonese (口语) are not rare in Cantonese songs since the early 1970s although formal written Chinese (书面语) is the norm. 黄霑 and 许冠杰 wrote many lyrics in colloquial Cantonese. 玩下啦,鬼马双星 and 半斤百两 are just some of the examples. Colloquial Cantonese are mostly used when the melodies are more playful or when the lyrics talk about very local topics and phenonmena. The more important issue is how the characters sound in a song. 嘅 or 唔 may not sound “right” or melodic. Older Cantonese lyrics don’t use that many 的 anyway. 黄霑 wrote many lyrics without using a single 的. He actually followed the style of classical literature. 的 is a modern invention in Chinese language. Check this out: http://stars.zaobao.com/pages1/huangzhan251104a.html.
On the classical, “high art” side, the consideration of tones is highly important even in the Mandarin-based 京剧 (jīngjù) and 昆曲 (kūnqǔ) operas. But traditionally, the tones are very much based on poetic 平仄 (píngzè) tone patterns rather than the actual contour tones of any Mandarin dialect (stage or otherwise), though that has been changing.
The status of melody in Chinese opera is quite interesting, as it is allowed to be changed within an intricate system of musical constraints to fit the tones. The traditional melodies as set by the scriptwriter are meant to be modified and ornamented upon in performance, including tone considerations. There’s a fascinating description and analysis of this phenomenon in Cantonese opera 粤剧 (yuèjù / yuhtkehk, jyut6kek6) at http://www.jstor.org/pss/850881
Bell Yung, Creative Process in Cantonese Opera I: The Role of Linguistic Tones, Ethnomusicology, Vol. 27, No. 1 (Jan., 1983), pp. 29-47.
In singing environments where melody is not as flexible, it can be a bit more… awkward. Hymns are a particular difficulty – the potential for “re-interpretation” of the text is huge, and the knowledge of what is meant does not necessarily alleviate any of the awkwardness. In Mandarin, I tend towards keeping the 2nd tone as rising with a grace note or a glissando; and I occasionally keep the 4th tone quite short, or add a glissando to keep its falling contour. In Cantonese… I sing in English (though I suppose on combining the above with the right transposition to fit the harmony at any one time…)!
I would assume that treatment of Western opera sung in Chinese is the same as that of Chinese pop songs.
is it significant that the Mandarin example used a foreign song? Is there a difference between tones of native and imported songs?
As a former music educator and current Chinese language teacher who often uses songs in teaching Chinese, I have thought about these issues.
When I participated in teacher training at Beijing Language and Culture University, I remember being taught that the different tones in mandarin have different durations. For example, the fourth tone lasts the shortest amount of time and the third tone lasts the longest. Usually we think of tones primarily in relationship to changes in pitch, not to duration. Duration of a syllable will obviously have a direct effect on the rhythm of both spoken and sung phrases. However, the effect is not on the pitches of the melody. At least, not in the Chinese folk, pop and children’s songs with which I am familiar.
I have been composing rhythmic chants for my students to help them remember key phrases of Chinese. It has been very clear that some rhythms “work” better than others because they are closer to the rhythm with which the same phrase would be spoken.
I tell my students that there are no tones in sung Mandarin and I still think that is the most useful thing I can tell them in a Chinese 1 or Chinese 2 class. But, what I really mean is that the PITCHES of Chinese tones are not present in sung Mandarin. What I do not mention, but think is true, is that the duration element of Mandarin tones does have a presence in the rhythms of sung Mandarin.
[…] own John Pasden actually wrote about this a few months back, coming to the conclusion that for Mandarin Chinese (as opposed to Cantonese), tones are more or […]
It is simply incorrect to say Mandarin songs ignore tones and only let melodies dominate. Those aren’t considered professional songs and end up sounding odd.
Some in the new generation have started slurring their words in song in an attempt to decrease the effect of the tone; however, this mostly sounds bad.
[…] I’ve found some suggestions that Cantonese actually tries harder to maintain relative tones within the tune, for example here: https://www.sinosplice.com/life/archives/2010/12/06/tones-in-chinese-songs. […]
[…] I’ve found some suggestions that Cantonese actually tries harder to maintain relative tones within the tune, for example here: https://www.sinosplice.com/life/archives/2010/12/06/tones-in-chinese-songs. […]
As I read the comments, I notice that almost everybody blames the poor lyricist for writing a text that doesn’t follow the contour of the melody, thus obscuring the tones and making it difficult to understand. I am a professional composer, and I can tell you that the normal procedure for writing a song is to FIRST write the lyrics and THEN set it to music. So, except in rare cases, the lyrics almost always come first, and the melody is composed to the lyrics to express the text. Therefore, if you hear a song where the tune doesn’t follow the word tones, you should first doubt the skills of the composer, and not of the lyricist. Of course, sometimes, if it is a well-known song written in another language, and then translated into Mandarin or Cantonese, it is then the job of the lyricist to make sure the tones match. As a composer, I can only say that it is much-much harder to write a good text to match a pre-existing tune than writing a tune on an existing text, and that’s true for any language, be it a tonal one or not!
Hi John,
I imagine you compose mostly in the Anglophone/European tradition, is that correct? The reason that most everyone is assuming that the music came first then text was set to the music is bc historically, that’s how songs were composed in China. Even the super famous Song Dynasty “ci” poems, which are now mostly recited as poetry, were originally written to match pre-existing music. To that end, Chinese actually has 4 separate terms to describe 1) composers of original music 作曲, 3) composers of music to match a pre-existing text 譜曲, 3) authors of original text 作詞, and 4) lyricists of text to match pre-existing music 填詞. The distinctions are starting to erode due to Western influence, but they exist bc we do have a cultural understanding of these differences.
We come from different musical traditions, and that’s cool, but it just means that you experience as a composer isn’t necessarily generalizable to this different historical and cultural context.
Updated url for “Tone and Melody in Cantonese” (Marjorie K. M. Chan):
http://journals.linguisticsociety.org/proceedings/index.php/BLS/article/view/1828
(doi [document object identifier]: 10.3765/bls.v13i0.1828)