How does tone develop?

[Aurora Rossa]

What are some ways in which tone can develop?

[pharazon]

I don't know much about this, but:

One way tones can originate is for voiced consonants to induce the next vowel to be pronounced in a low pitch.

[alice]

What are some ways in which tone can develop?

In earlier stages of the Slavonic languages, long vowels could develop from Proto-Indo-European long vowels or diphthongs. Those from long vowels had a rising tone (normally known as "acute"), those from diphthongs had a falling tone ("circumflex"). Thus PIE /e:/ > Slavonic /e:/ with rising tone, whereas PIE /ai/ > /e:/ with falling tone.

Another rising tone ("neoacute") developed on short vowels which were lengthened after the loss of short /i u/ in the following syllable. I can't think of an actual example offhand; something like /katu/ would become /ka:t/, where the /a:/ would have the neoacute. The subsequent development of the tones was complicated.

[jburke]

What are some ways in which tone can develop?

In Cheyenne, contraction of Proto Algonquian long vowels gave rise to pitch; long vowels := to high-pitched vowels.

[Eddy the Great(NLI)]

How did tone develop in, say, Chinese or Vietnamese?

[Soap]

I think proto-Chinese had some pretty monstrous consonant clusters, and in nearly all dialects the consonants came to tonalize the preceding or following vowel. For example, voiced consonants caused the preceding vowel to be pronounced on a lower tone. Vietnamese was originally toneless as well, and acquired its tones by massive amounts of loaning from Chinese and other languages ... this caused some shifts in the native words as well, probably similar to the ones above.

[Aurora Rossa]

In earlier stages of the Slavonic languages, long vowels could develop from Proto-Indo-European long vowels or diphthongs. Those from long vowels had a rising tone (normally known as "acute"), those from diphthongs had a falling tone ("circumflex"). Thus PIE /e:/ > Slavonic /e:/ with rising tone, whereas PIE /ai/ > /e:/ with falling tone.

Another rising tone ("neoacute") developed on short vowels which were lengthened after the loss of short /i u/ in the following syllable. I can't think of an actual example offhand; something like /katu/ would become /ka:t/, where the /a:/ would have the neoacute. The subsequent development of the tones was complicated.

Slavonic langs are tonal?

How might tone develop in my conlang? Could words like !'?la and tqus?(where ^ indicates pharyngealization) develop tones? How?

[pharazon]

Slavonic langs are tonal?

Modern Slavic languages aren't.

[Soap]

Somehow, tonal languages developed in the isolated South Pacific in the absence of any other tonal languages, in languages without any serious consonant clusters and only 5 basic vowels. I don't know how it happened, but it seems to me that there must be a lot of different ways to get to having a fully tonal language. Also, proto-IE was slightly tonal (pitch accent is the correct term for it), and that must have happened somehow too, since as far as we know IE isnt related to any tone languages other than a few modern IE ones.

[Aurora Rossa]

There is no voicing contrast so it can't develop that way and there is no vowel length contrast so it can't develop that way. How else can tone develop or is those the only ways?

[Glenn]

Slavonic langs are tonal?

Modern Slavic languages aren't.

According to Wikipedia (here), modern Serbo-Croatian (Serbian/Croatian/Bosnian) does have a "melodic accent", apparently involving falling and rising tones related to stressed syllables; the same term is used to describe (in more detail) the intonation patterns of Swedish and Norwegian.

I've long been aware of the "bouncy" intonation of Swedish (parodied here in the U.S. by the character of the Swedish Chef on the old Muppet Show), but until recently, I never thought about it as a tonal phenomenon with its own sound rules--which, however, it is.

p@,
Glenn

[ran not in]

There is no voicing contrast so it can't develop that way and there is no vowel length contrast so it can't develop that way. How else can tone develop or is those the only ways?

Final consonants... Many of the tonal distinctions in Chinese and Vietnamese can be dated back to lost final consonants like -s, -?, -h, etc. "-s" was, in particular, a suffix of some sort in Old Chinese, so that it created a lot of doublets differing in only tone and related in meaning.

You can find this on the internet btw. I remember seeing a pdf file on Vietnamese tonogenesis.. you can look that up on Google if you want.

[Aurora Rossa]

What effect would pharyngealization have on tone? Could it evolve into tone?

[orkybash]

I don't see why it couldn't. Seems plausible to me to have pharangyalized vowels take a low tone and non-pharangyalized vowels take a high tone...

[Aurora Rossa]

I don't see why it couldn't. Seems plausible to me to have pharangyalized vowels take a low tone and non-pharangyalized vowels take a high tone...

That's what I thought. Pharyngeal vowels are definantly deeper or darker sounding.

[pharazon]

According to Wikipedia (here), modern Serbo-Croatian (Serbian/Croatian/Bosnian) does have a "melodic accent", apparently involving falling and rising tones related to stressed syllables; the same term is used to describe (in more detail) the intonation patterns of Swedish and Norwegian.

Yeah, I remember about Serbo-Croatian now; I was thinking just of purely tonal languages though, rather than pitch-accent ones (although there is overlap). On this note, do pitch-accent systems come about in the same ways as tonal systems like Mandarin?

[alice]

Slavonic langs are tonal?

Modern Slavic languages aren't.

Except for Serbo-Croat and Slovene. Older forms of Slavonic were, too; the tonal distinctions are reflected today in vowel length (Czech, Slovak, and partly in Polish, Polabian, Sorbian and Cassubian) or place of stress (Russian, Ukranian, Belorussian, Bulgarian, and Macedonian (I think)).

[alice]

In earlier stages of the Slavonic languages, long vowels could develop from Proto-Indo-European long vowels or diphthongs.

I unwittingly used something similar in one of my conlangs. In this conlang a vowel could be preceded or followed by a glide; the glides had a low pitch and the vowels a high pitch, so a different tone developed from each of the possibilities:

- Vowel only -> level (high) tone (a)
- Glide + vowel -> rising tone (?)
- Vowel + glide -> falling tone (?)
- Glide + vowel + glide -> rising-falling tone (?)

The tones only appeared on the vowel in the initial syllable, which bore the main stress, and they became phonemic when the glides later disappeared. There was also a prohibition on a syllable starting and ending on a low pitch, which meant that in polysyllables, a falling tone or the falling part of the rising-falling tone was pushed to the final syllable:

? a -> a ?
? a -> ? ?

Does this sort of thing exist anywhere else, or is it unique to this conlang?

[Aurora Rossa]

I had another idea. Perhaps vowels after certain final consonants could be dropped and transfer their tone to the preceeding vowel, making a rising or falling tone. Note that ?? indicates a rising tone.

!'?l? --> !'??l

Also, should both tones(high and low) be indicated by diacritics?

[olsem nakato]

I've long been aware of the "bouncy" intonation of Swedish (parodied here in the U.S. by the character of the Swedish Chef on the old Muppet Show), but until recently, I never thought about it as a tonal phenomenon with its own sound rules--which, however, it is.

It might have rules, but it has no phonemic significance. The Swedish of Finland doesn't bounce, it goes like a train down a slope; yet everybody understands what is being said. (The Swedish of Sweden, Rikssvenska, got its bouncy intonation just a hundred or so years ago. The pronunciation of Swedish before that was rather flat and descending, as it still is in Finland. Swedish got to sound like that because the Swedes one day woke up and simultaneously and unanimously decided they wanted to talk fancy. Or so I've been told. )

[gach]

I've long been aware of the "bouncy" intonation of Swedish (parodied here in the U.S. by the character of the Swedish Chef on the old Muppet Show), but until recently, I never thought about it as a tonal phenomenon with its own sound rules--which, however, it is.

It might have rules, but it has no phonemic significance. The Swedish of Finland doesn't bounce, it goes like a train down a slope; yet everybody understands what is being said. (The Swedish of Sweden, Rikssvenska, got its bouncy intonation just a hundred or so years ago. The pronunciation of Swedish before that was rather flat and descending, as it still is in Finland. Swedish got to sound like that because the Swedes one day woke up and simultaneously and unanimously decided they wanted to talk fancy. Or so I've been told. )

A Swedish analogue discribes Finnish and finnish Swedish as kanalb?t (canal boat) and Riksvenska as segelb?t (sail boat). It describes them pretty well.

[vilk]

ran not in had it right, but there's more: tone contour, at least in Chinese, Tibetan, Vietnamese, and most likely other SE Asian languages, developed from the final consonant and the tone register (IE, high pitch or low pitch) from the initial consonant. Given an opposition in voice in a syllable-initial consonant, the unvoiced (I think) yields a high initial pitch and the voiced a low initial pitch. From there, a syllable-final [?] would raise the pitch and a [h] would lower it. Of course, these two weren't always the only codae available but by the time tone developed, it is theorized, most endings had become [?] or [h] through lenition. (I may be wrong, as I'm gleaning most of this information off the web.)

Andr?-Georges Haudricourt put forward his theory of tonogenesis in 1954; previously it was thought that tonal and non-tonal languages could not be genetically related. His theory was bolstered by a study conducted by Jean-Marie Hombert in 1978 in which he recorded the change in fundamental tone in vowels preceding syllable-final [?] and [h], as pronounced by native speakers of Arabic. It was found that vowels preceding [?] went up in pitch at least 25 Hz, while vowels preceding [h] went down in pitch at least 9 Hz. Voila. With an initial pitch either high or low and pitch movement dictated by the coda, you have tone.

As for tone sandhi: I have no idea, but it seems likely that ease of articulation is the culprit.

[Ran]

ran not in had it right, but there's more: tone contour, at least in Chinese, Tibetan, Vietnamese, and most likely other SE Asian languages, developed from the final consonant and the tone register (IE, high pitch or low pitch) from the initial consonant. Given an opposition in voice in a syllable-initial consonant, the unvoiced (I think) yields a high initial pitch and the voiced a low initial pitch. From there, a syllable-final [?] would raise the pitch and a [h] would lower it. Of course, these two weren't always the only codae available but by the time tone developed, it is theorized, most endings had become [?] or [h] through lenition. (I may be wrong, as I'm gleaning most of this information off the web.)

Andr?-Georges Haudricourt put forward his theory of tonogenesis in 1954; previously it was thought that tonal and non-tonal languages could not be genetically related. His theory was bolstered by a study conducted by Jean-Marie Hombert in 1978 in which he recorded the change in fundamental tone in vowels preceding syllable-final [?] and [h], as pronounced by native speakers of Arabic. It was found that vowels preceding [?] went up in pitch at least 25 Hz, while vowels preceding [h] went down in pitch at least 9 Hz. Voila. With an initial pitch either high or low and pitch movement dictated by the coda, you have tone.

As for tone sandhi: I have no idea, but it seems likely that ease of articulation is the culprit.

Yep - but remember that the distinction between tone contour and tone register does not have to be reflected in the resulting tone categories of a language. In Chinese, for example, the "ideal" system would be:

CONTOURS - ping (flat) - shang (rising) - qu (falling)
(from no final, -? final, and -h final)
REGISTERS - yin (high) - yang (low)
(from voiceless initial and voiced initial)

But few dialects actually make all the six-way distinctions. Mandarin, for example, has merged yang-shang, yin-qu, and yang-qu (or never split the qu according to register in the first place), and in fact the modern yang-shang (together with qu) is HIGHER than yin-qu in pitch; in effect the tones are distinguished through a combination of contour and register.

In any case the actual values of the "contours" are themselves in doubt - they may very well have been mid/high/low, instead of flat/rise/fall.

[Aurora Rossa]

How might tone sandhi develop?

[pharazon]

It develops just like other sandhi; certain combinations of sounds (tones, in this case) are more difficult to say than others, so they change.