Tonogenesis

A forum for all topics related to constructed languages
Post Reply
User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Tonogenesis

Post by Shemtov » 19 Jun 2017 01:43

I have a language that has a two-tone system: /˧ ˥/. It has voicing contrast and word final voiceless stops. I want to, in one descendant, unvoice the voiced stops and fricatives, producing ˥>˧ and ˧>˩ and to delete word final stops causing, depending on whether the original consonant was voiced or not /˦˥ ˩˨/ to enter the tonemes. However, there's one descendant whom I want to devoice the voiced stops and fricatives, but keep the word final stops. Is there any way that this descendants tonemes will be contour; i.e. beyond /˥ ˧ ˩/?
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
sangi39
moderator
moderator
Posts: 3356
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: Tonogenesis

Post by sangi39 » 19 Jun 2017 01:57

Shemtov wrote:I have a language that has a two-tone system: /˧ ˥/. It has voicing contrast and word final voiceless stops. I want to, in one descendant, unvoice the voiced stops and fricatives, producing ˥>˧ and ˧>˩ and to delete word final stops causing, depending on whether the original consonant was voiced or not /˦˥ ˩˨/ to enter the tonemes. However, there's one descendant whom I want to devoice the voiced stops and fricatives, but keep the word final stops. Is there any way that this descendants tonemes will be contour; i.e. beyond /˥ ˧ ˩/?
So, just to throw a couple of examples together:

*ta˧ > *ta˧ > *ta˧
*ta˥ > *ta˥ > *ta˧
*da˧ > *da˩ > *ta˧
*da˥ > *da˧ > *ta˧
*tap˧ > *tap˧ > *tap˦˥ > *tap˦˥ > *ta˦˥
*tap˥ > *tap˥ > *tap˦˥ > *tap˦˥ > *ta˦˥
*dap˧ > *dap˩ > *dap˩˨ > *tap˩˨ > *ta˩˨
*dap˥ > *dap˧ > *dap˦˥ > *tap˦˥ > *ta˦˥

?

I'd say that it's possible that in the language that keeps final plosives the tones might be contour tones, but they might just be allophonic realisations of the other tones in closed syllables.

Are long vowels phonemic in the original language?
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Re: Tonogenesis

Post by Shemtov » 19 Jun 2017 02:01

*ta˧ > *ta˧ > *ta˧
*ta˥ > *ta˥ > *ta˥
*da˧ > *da˩ > *ta˩
*da˥ > *da˧ > *ta˧
*tap˧ > *tap˧ > *tap˦˥ > *tap˦˥ > *ta˦˥
*dap˧ > *dap˩ > *dap˩˨ > *tap˩˨ > *ta˩˨




sangi39 wrote:
Are long vowels phonemic in the original language?
No.
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
sangi39
moderator
moderator
Posts: 3356
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: Tonogenesis

Post by sangi39 » 19 Jun 2017 02:31

Shemtov wrote:*ta˧ > *ta˧ > *ta˧
*ta˥ > *ta˥ > *ta˥
*da˧ > *da˩ > *ta˩
*da˥ > *da˧ > *ta˧
*tap˧ > *tap˧ > *tap˦˥ > *tap˦˥ > *ta˦˥
*dap˧ > *dap˩ > *dap˩˨ > *tap˩˨ > *ta˩˨
Ohhh, so there was just a mid tone if the syllable was closed.

Well, you could have the initial consonant only exert partial tone on the syllable's tone as a whole:

*ta˧ > *ta˥˧ > *ta˥˧
*ta˥ > *ta˥ > *ta˥
*da˧ > *da˧ > *ta˧
*da˥ > *da˧˥ > *ta˧˥
*tap˧ > *tap˧˥ > *tap˧˥ > *tap˧˥
*dap˧ > *dap˧ > *dap˧ > *tap˧

The tones in open syllables could shift, since the distinction is "higher", "lower", "rising" and "falling" with closed syllables having "level" and "rising" tones.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
Creyeditor
mongolian
mongolian
Posts: 4543
Joined: 14 Aug 2012 19:32

Re: Tonogenesis

Post by Creyeditor » 19 Jun 2017 02:38

Shemtov wrote:I have a language that has a two-tone system: /˧ ˥/. It has voicing contrast and word final voiceless stops. I want to, in one descendant, unvoice the voiced stops and fricatives, producing ˥>˧ and ˧>˩ and to delete word final stops causing, depending on whether the original consonant was voiced or not /˦˥ ˩˨/ to enter the tonemes. However, there's one descendant whom I want to devoice the voiced stops and fricatives, but keep the word final stops. Is there any way that this descendants tonemes will be contour; i.e. beyond /˥ ˧ ˩/?
*ta˧ > ta˥˧ > ta˥˧ (> ta˧)
*ta˥ > ta˥ > ta˥
*da˧ > da˩˧ > ta˩˧ ( > ta˩)
*da˥ > da˩˥ > ta˩˥
*tap˧ > tap˥˧ > tap˥˧ (> ta˧)
*dap˧ > dap˩˧ > dap˩˨ ( > ta˩)

This is my take on this. Voicing causes the onset of the vowels pitch to lower, whereas voicelessness raises it. This gives you high-falling, high level, low rising and rising tone. These could of course simplify, e.g. to mid-level, high-level, low level and rising, but I think staying contour tones would also be okay. You can keep the codas and still get at least one contour. If the ancestor language has stress, you could get even more tones.

stressed
*ˈta˧ > ta˥˧ > ta˥˧
*ˈta˥ > ta˥ > ta˥
*ˈda˧ > da˩˧ > ta˩˧
*ˈda˥ > da˩˥ > ta˩˥
*ˈtap˧ > tap˥˧ > tap˥˧
*ˈdap˧ > dap˩˧ > dap˩˨

unstressed
*ta˧ > ta˩ > ta˥˩ > ta˥˩
*ta˥ > ta˧ > ta˥˧ > ta˥˧
*da˧ > da˩ > da˩ > ta˩
*da˥ > da˧ > da˩˧ > ta˩˧
*tap˧ >tap˩ > tap˥˩ > tap˥˩
*dap˧ >dap˩ > dap˩ > dap˩

This would give you a system with two level tones: high and low and several contour tones: high falling, falling, rising and low rising. Of course, one tone might simplify into a mid tone, but you would still get three contours.
Edit: I feel a bit ninja'd [:D]
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]

User avatar
Frislander
runic
runic
Posts: 3498
Joined: 14 May 2016 18:47
Location: The North

Re: Tonogenesis

Post by Frislander » 19 Jun 2017 11:56

If you want a good look at how systems like these can develop then the Wikipedia article on the four tones of Middle Chinese is great, it has an absolutely massive table of their varying outcomes in the different Sinitic topolects.

User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Re: Tonogenesis

Post by Shemtov » 19 Jun 2017 12:30

Creyeditor wrote:
Shemtov wrote:I have a language that has a two-tone system: /˧ ˥/. It has voicing contrast and word final voiceless stops. I want to, in one descendant, unvoice the voiced stops and fricatives, producing ˥>˧ and ˧>˩ and to delete word final stops causing, depending on whether the original consonant was voiced or not /˦˥ ˩˨/ to enter the tonemes. However, there's one descendant whom I want to devoice the voiced stops and fricatives, but keep the word final stops. Is there any way that this descendants tonemes will be contour; i.e. beyond /˥ ˧ ˩/?
*ta˧ > ta˥˧ > ta˥˧ (> ta˧)
*ta˥ > ta˥ > ta˥
*da˧ > da˩˧ > ta˩˧ ( > ta˩)
*da˥ > da˩˥ > ta˩˥
*tap˧ > tap˥˧ > tap˥˧ (> ta˧)
*dap˧ > dap˩˧ > dap˩˨ ( > ta˩)

This is my take on this. Voicing causes the onset of the vowels pitch to lower, whereas voicelessness raises it. This gives you high-falling, high level, low rising and rising tone. These could of course simplify, e.g. to mid-level, high-level, low level and rising, but I think staying contour tones would also be okay. You can keep the codas and still get at least one contour. If the ancestor language has stress, you could get even more tones.

stressed
*ˈta˧ > ta˥˧ > ta˥˧
*ˈta˥ > ta˥ > ta˥
*ˈda˧ > da˩˧ > ta˩˧
*ˈda˥ > da˩˥ > ta˩˥
*ˈtap˧ > tap˥˧ > tap˥˧
*ˈdap˧ > dap˩˧ > dap˩˨

unstressed
*ta˧ > ta˩ > ta˥˩ > ta˥˩
*ta˥ > ta˧ > ta˥˧ > ta˥˧
*da˧ > da˩ > da˩ > ta˩
*da˥ > da˧ > da˩˧ > ta˩˧
*tap˧ >tap˩ > tap˥˩ > tap˥˩
*dap˧ >dap˩ > dap˩ > dap˩

This would give you a system with two level tones: high and low and several contour tones: high falling, falling, rising and low rising. Of course, one tone might simplify into a mid tone, but you would still get three contours.
Edit: I feel a bit ninja'd [:D]
Are you saying that there would be contour tones that would only apply to stop-final syllables?
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
Creyeditor
mongolian
mongolian
Posts: 4543
Joined: 14 Aug 2012 19:32

Re: Tonogenesis

Post by Creyeditor » 19 Jun 2017 12:55

No, the tones are independent of the coda in my suggestion.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]

TwistedOne151
hieroglyphic
hieroglyphic
Posts: 41
Joined: 11 Nov 2012 19:52

Re: Tonogenesis

Post by TwistedOne151 » 22 Jun 2017 12:51

Frislander wrote:If you want a good look at how systems like these can develop then the Wikipedia article on the four tones of Middle Chinese is great, it has an absolutely massive table of their varying outcomes in the different Sinitic topolects.
Yeah, I looked at that table, along with the comparisons of the forms of the six (or five, depending on dialect) tones of Vietnamese in the different regions here, this paper (PDF) on the outcomes for different Tai Khuen varieties of the tone split in the four Proto-Tai tones, this paper (PDF download) on some of the tone systems from the tone split of Proto-Tai tones in the Zhuang languages, (as well as comparing standard Thai, Lao, and Bouyei tones, and looking at tone system outcomes in the Hmong-Mien languages from various sources) for my own work on a tonal language for trying to figure out the outcome of a tone split.

(Mainly, I found it hard to find any sort of consistent pattern to the outcomes; I'll admit I'm still not sure how to determine if an outcome for a tone split is plausible or not.)

User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Re: Tonogenesis

Post by Shemtov » 25 Jun 2017 04:46

I'm thinking of adding vowel length in a transitional state of the language and then erasing it. How would that effect tone?
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
sangi39
moderator
moderator
Posts: 3356
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: Tonogenesis

Post by sangi39 » 25 Jun 2017 05:11

Shemtov wrote:I'm thinking of adding vowel length in a transitional state of the language and then erasing it. How would that effect tone?
From what I've been able to gather here and there, long vowels, precisely because they have a prolonged duration, are capable of distinguishing a higher number of tones, especially contour tones, than short vowels. It seems to be a case, basically, of because the vowel is held longer, there's more time in which a tone has to be produced, therefore there's more time for it to be distinguished (from other level tones), rise, fall, etc.

Having long vowels that can hold tone, for example, might make it more likely to see contour tones appearing on short vowels when vowel length is dropped (note, however, that "vowel length" is a relative term, not an absolute one, so long vowels might shorten, but short vowels might simultaneously lengthen, such that long and short vowels merge into a short of "medium" length, but since vowel length is no longer distinguished, that's just how long vowels are held for, and this could even be conditioned by stress or syllable openness).

At the end of the day, long vowels probably give you more room to work with what tones you can have once vowel length becomes non-contrastive.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Re: Tonogenesis

Post by Shemtov » 25 Jun 2017 05:22

sangi39 wrote:
Shemtov wrote:I'm thinking of adding vowel length in a transitional state of the language and then erasing it. How would that effect tone?
From what I've been able to gather here and there, long vowels, precisely because they have a prolonged duration, are capable of distinguishing a higher number of tones, especially contour tones, than short vowels. It seems to be a case, basically, of because the vowel is held longer, there's more time in which a tone has to be produced, therefore there's more time for it to be distinguished (from other level tones), rise, fall, etc.

Having long vowels that can hold tone, for example, might make it more likely to see contour tones appearing on short vowels when vowel length is dropped (note, however, that "vowel length" is a relative term, not an absolute one, so long vowels might shorten, but short vowels might simultaneously lengthen, such that long and short vowels merge into a short of "medium" length, but since vowel length is no longer distinguished, that's just how long vowels are held for, and this could even be conditioned by stress or syllable openness).

At the end of the day, long vowels probably give you more room to work with what tones you can have once vowel length becomes non-contrastive.
I'm asking what kind of tones can devolop from the fusion of long and short vowels? Rising? Falling? Dipping? Would the contours be on the formerly long vowels or the short vowels? Or both, but different?
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
Frislander
runic
runic
Posts: 3498
Joined: 14 May 2016 18:47
Location: The North

Re: Tonogenesis

Post by Frislander » 25 Jun 2017 13:19

Shemtov wrote:
Spoiler:
sangi39 wrote:
Shemtov wrote:I'm thinking of adding vowel length in a transitional state of the language and then erasing it. How would that effect tone?
From what I've been able to gather here and there, long vowels, precisely because they have a prolonged duration, are capable of distinguishing a higher number of tones, especially contour tones, than short vowels. It seems to be a case, basically, of because the vowel is held longer, there's more time in which a tone has to be produced, therefore there's more time for it to be distinguished (from other level tones), rise, fall, etc.

Having long vowels that can hold tone, for example, might make it more likely to see contour tones appearing on short vowels when vowel length is dropped (note, however, that "vowel length" is a relative term, not an absolute one, so long vowels might shorten, but short vowels might simultaneously lengthen, such that long and short vowels merge into a short of "medium" length, but since vowel length is no longer distinguished, that's just how long vowels are held for, and this could even be conditioned by stress or syllable openness).

At the end of the day, long vowels probably give you more room to work with what tones you can have once vowel length becomes non-contrastive.
I'm asking what kind of tones can devolop from the fusion of long and short vowels? Rising? Falling? Dipping? Would the contours be on the formerly long vowels or the short vowels? Or both, but different?
You'd get the contours (of any kind) on the long vowels. The main idea behind languages distinguishing contour tones on long vowels but only level tones on short ones (for example Navajo) is that you analyse the long vowels as underlyingly being sequences of short vowels with their corresponding level tones (so for instance a high and a low vowel becomes a fall whereas a low and a high becomes a rise, with only sequences of high plus high giving long high vowels). These analyses are supported by the fact that diphthongs behave the same way as normal long vowels do. Once they merge with the short vowels, though, all kinds of shifts may be possible, as this analysis of vowel sequences is no longer possible, making the contour tones properly phonemic.

User avatar
Shemtov
runic
runic
Posts: 3138
Joined: 29 Apr 2013 04:06

Re: Tonogenesis

Post by Shemtov » 25 Jun 2017 13:27

Can POA of the coda affect tone? I'm thinking of making coda /m/>/n/ and coda /p/ >/k/ (in fact, for a Period the language will use phonemic /p/ entirely). Would that affect tone at all?
Many children make up, or begin to make up, imaginary languages. I have been at it since I could write.
-JRR Tolkien

User avatar
Frislander
runic
runic
Posts: 3498
Joined: 14 May 2016 18:47
Location: The North

Re: Tonogenesis

Post by Frislander » 25 Jun 2017 13:46

Shemtov wrote:Can POA of the coda affect tone? I'm thinking of making coda /m/>/n/ and coda /p/ >/k/ (in fact, for a Period the language will use phonemic /p/ entirely). Would that affect tone at all?
No I don't think so unless you allow glottals in the coda, which might behave differently to other stops (e.g. a glottal stop might leave creaky voice while other stops leave a high tone). On the other hand the POA could affect the rounding/frontedness of the vowel (e.g. coronals might cause a back vowel vowel to front, a in Tibetan).

User avatar
sangi39
moderator
moderator
Posts: 3356
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: Tonogenesis

Post by sangi39 » 26 Jun 2017 01:21

Frislander wrote:
Shemtov wrote:
Spoiler:
sangi39 wrote:
Shemtov wrote:I'm thinking of adding vowel length in a transitional state of the language and then erasing it. How would that effect tone?
From what I've been able to gather here and there, long vowels, precisely because they have a prolonged duration, are capable of distinguishing a higher number of tones, especially contour tones, than short vowels. It seems to be a case, basically, of because the vowel is held longer, there's more time in which a tone has to be produced, therefore there's more time for it to be distinguished (from other level tones), rise, fall, etc.

Having long vowels that can hold tone, for example, might make it more likely to see contour tones appearing on short vowels when vowel length is dropped (note, however, that "vowel length" is a relative term, not an absolute one, so long vowels might shorten, but short vowels might simultaneously lengthen, such that long and short vowels merge into a short of "medium" length, but since vowel length is no longer distinguished, that's just how long vowels are held for, and this could even be conditioned by stress or syllable openness).

At the end of the day, long vowels probably give you more room to work with what tones you can have once vowel length becomes non-contrastive.
I'm asking what kind of tones can devolop from the fusion of long and short vowels? Rising? Falling? Dipping? Would the contours be on the formerly long vowels or the short vowels? Or both, but different?
You'd get the contours (of any kind) on the long vowels. The main idea behind languages distinguishing contour tones on long vowels but only level tones on short ones (for example Navajo) is that you analyse the long vowels as underlyingly being sequences of short vowels with their corresponding level tones (so for instance a high and a low vowel becomes a fall whereas a low and a high becomes a rise, with only sequences of high plus high giving long high vowels). These analyses are supported by the fact that diphthongs behave the same way as normal long vowels do. Once they merge with the short vowels, though, all kinds of shifts may be possible, as this analysis of vowel sequences is no longer possible, making the contour tones properly phonemic.
Frislander put it much better there than I did [:)] Thanks for the example as well. The best I could do was the pitch accent of Ancient Greek.

I'd assume that when vowel length became non-constrastive, it could depend on exactly how it became non-contrastive, e.g. is it dependent on stress, syllable openness, a combination of the two? If all long vowels shortened to the length of original short vowels then I'd imagine that the number of possible contour tones would decrease (either merging into each other or some contours become level), but if short vowels become "long" in some syllables at the same time long vowels became "short" in others, then contour tones might collapse in some syllables while remaining distinct in others.


Frislander wrote:
Shemtov wrote:Can POA of the coda affect tone? I'm thinking of making coda /m/>/n/ and coda /p/ >/k/ (in fact, for a Period the language will use phonemic /p/ entirely). Would that affect tone at all?
No I don't think so unless you allow glottals in the coda, which might behave differently to other stops (e.g. a glottal stop might leave creaky voice while other stops leave a high tone). On the other hand the POA could affect the rounding/frontedness of the vowel (e.g. coronals might cause a back vowel vowel to front, a in Tibetan).
I'd agree with you on this one. The main consonants that seem to affect tone differently from consonants with the same MOA but at different POAs are the glottals and the pharyngeals. Similarly, certain phonation types seem to affect tone as well, and in some languages they're tied together, e.g. in Burmese.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

User avatar
Creyeditor
mongolian
mongolian
Posts: 4543
Joined: 14 Aug 2012 19:32

Re: Tonogenesis

Post by Creyeditor » 26 Jun 2017 13:30

Shemtov wrote:Can POA of the coda affect tone? I'm thinking of making coda /m/>/n/ and coda /p/ >/k/ (in fact, for a Period the language will use phonemic /p/ entirely). Would that affect tone at all?
One way that tone could be affected indirectly is if you have other changes interacting with it. If coronal consonants are deleted at the end of a syllable and this then causes a chain shift as you described. If tonogenesis is now depended on closed and open syllables, this could cause a difference.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]

Post Reply