Linguistic tone and song in Indo-Aryan

Joan Baart joanbaart at HOTMAIL.COM
Sat Apr 24 07:00:35 UTC 2004

VYAKARAN: South Asian Languages and Linguistics Net
Editors:  Tej K. Bhatia, Syracuse University, New York
          John Peterson, University of Osnabrueck, Germany
Details:  Send email to listserv at and say: INFO VYAKARAN
Subscribe:Send email to listserv at and say:
          (Substitute your real name for first_name last_name)

Dear Peter and all,

First of all, I conclude from the lack of response to my query that nobody among us is aware of any studies about the relation of linguistic tone and song in Indo-Aryan languages, or even more generally South Asian languages. This is kind of what I was expecting.

It is interesting that you bring up Chinese, Peter. There is a recent article by Patrick Wong and Randy Diehl on tone and song in Chinese. It can be retrieved online from

Here is a quick summary of their findings: Chao (1956) investigated the relationship between sung pitch and linguistic tone in Chinese songs of various styles. In Chinese "Singsong" (a style that is intermediate between speaking and singing), each tone is sung with a consistent pitch pattern, which makes it relatively easy for listeners to identify tones and, hence, word meanings. On the other hand, in contemporary Mandarin songs, composers mostly ignore linguistic tones in their compositions, according to Chao. Yung (1983) looked at Chinese opera and found a relatively consistent relation between melody and tone, comparable to what Chao found for Singsongs. In their article, Wong & Diehl themselves analyzed four contemporary Cantonese songs. They looked at direction of pitch change over pairs of consecutive syllables and found an overall correspondence of over 90 percent between musical and tonal sequences. So, while the fundamental frequency intervals and the shape of the contours that are normal for speech are not reproduced exactly in these songs, there does seem to be a very strong tendency for a rising sequence of tones to correspond with an ascending sequence of musical notes, and for a falling sequence of tones to correspond with a descending sequence of notes, at least in the style that Wong & Diehl looked at.

This week I have been applying this methodology to a sample of Kalam Kohistani songs, and found that in about two-thirds (66 percent) of the instances in my sample, a rising sequence of tones corresponds to rising pitch in the songs, and a falling sequence of tones to falling pitch in the songs. This is a much weaker correspondence as compared to Cantonese (66 percent vs. 90 percent), which could be due to a number of factors, but one of these is probably that the songs I have been looking at all use a fixed, traditional tune. The tune is, as it were, imposed on the lyrics, rather than newly composed to go with the lyrics.

There is unpublished work by colleagues of mine who have looked at Thai songs of various styles. They found the highest correspondence (around 90 percent) of tones and sung pitches in some classical and traditional Thai songs. For contemporary popular songs (that often borrow elements from western popular music) the number goes down to 60-70 percent. For a western hymn, translated into Thai, the number of matching correspondences was 42 percent. The worst case is the national anthem, with a correspondence of only 32 percent. This result confirms the notion, I think, that the degree to which a melody is "imposed" influences the likelihood for tones and melody to match. (The Thai national anthem used a pre-existing traditional melody to which new lyrics were set.)

Anyway, what I have been learning so far is that there is a wide range of variation as to tone-melody matching across languages, and often also within languages.



----- Original Message ----- 
From: Peter Hook 
Sent: Monday, April 19, 2004 10:07 PM
Subject: Linguistic tone and song in Indo-Aryan

Dear Joan,
        You may find it useful to look at the relation between pitches and
tones in the Chinese "ci". Most of the "ci" melodies have been lost but
there are 4 or 5 from the Sung Dynasty that were somehow preserved.
        However, I'm not optimistic that you will find any relationship.
While some classical Chinese prosodies require the placement of words
having specified tones at certain points in a line of poetry, I have not
heard of lexical tone having any relation to the pitches in a melody. On
the contrary, in Chinese, tonal distinguishes disappear altogether when
one sings. I suspect that will turn out to be the case in Panjabi, SHina,
Kohistani, and other S Asian tone languages, too.

        All the best,


On Mon, 19 Apr 2004, Joan Baart wrote:

> Dear colleagues,
> I am working on a short paper on prosody and poetry in Kalam Kohistani (Pakistan). Kalam Kohistani is a tone language, and one of the questions I am asking is if there is a systematic relation in this language between the pitches of a song and the phonological tones of the words of the song. (It would seem that there isn't; at least not in the styles I have recorded.)
> I am wondering if this specific question has been addressed before for any Indo-Aryan language. One would think, for instance, that somebody must have looked at Punjabi tone and songs, but I have no references to any work in this area. The only material that I am aware of at the moment concerns Vedic accents and the extent to which they are preserved in recitation (Wayne Howard: "Vedic chant" in the Garland Encyclopedia of World Music, Vol. 5).
> If you are aware of any relevant work, please let me know.
> With best wishes,
> Joan Baart

Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (
Version: 6.0.656 / Virus Database: 421 - Release Date: 4/9/2004
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Vyakaran mailing list