zompist bboard

Zaarin wrote:I have a rather weird one I want to ask about. s/k/_# Does that seem too far fetched?

Seems unlikely.
Perhaps:
gliding [s→j/_#]
fortition [j→c/_#]
depalatalization [c→k/_#]
But I wouldn't expect fortition of [j] in final position.
Maybe derive it by morphology. Like if -s is an affix, -k starts taking over most of its roles instead?

Sexendèƚo wrote:

Zaarin wrote:I have a rather weird one I want to ask about. s/k/_# Does that seem too far fetched?

Seems unlikely.
Perhaps:
gliding [s→j/_#]
fortition [j→c/_#]
depalatalization [c→k/_#]
But I wouldn't expect fortition of [j] in final position.
Maybe derive it by morphology. Like if -s is an affix, -k starts taking over most of its roles instead?

Yeah, I'm working on a classical Indo-European language, and where most such languages (Latin, Greek, Gaulish, etc.) have s in their thematic nominal declensions, I want to have k. Because the endings are fusional, a morphological explanation would be difficult; even if I added an excrescent /k/ word-finally, /sk/ becomes /ʃ/ in this language. Hmm, what about s > x > k word-finally?

Or, even quicker, debuccalisation (s>h) or a similar process (s>x) followed by fortition (h>k or x>k).

jmcd wrote:Or, even quicker, debuccalisation (s>h) or a similar process (s>x) followed by fortition (h>k or x>k).

Thanks, I think that's the route I'll take.

How might back unrounded vowels diphthongize? I'm mostly familiar with diphthongs in English and Quebec French, and I'm wondering what we might expect [ɯː],[ɤː],[ʌː],[ɑː] to diphthongize to.

Frislander wrote:I'm back-deriving the proto-inventory for this language, and I have a phoneme /ʈʳ/ and I want it to split and caused a position-based alternation in the daughter language. Does changing it to /t/ word-initially and /t͡ʃ/ otherwise seem sensible?

I think this was buried, but my question still stands: does this seem OK?

Additional information: this change comes in the context of a group of changes related to fricatives which sees /ɸ/ become /p/ initially and /w/ otherwise, /ʃ/ becoming /s/ initially and /j/ otherwise, and /ʂ/ to /s/ initially and /ɽ/ otherwise.

Sexendèƚo wrote:How might back unrounded vowels diphthongize? I'm mostly familiar with diphthongs in English and Quebec French, and I'm wondering what we might expect [ɯː],[ɤː],[ʌː],[ɑː] to diphthongize to.

For the most part, analogously to their counterparts with the same closeness in other columns. For example, in Vietnamese close-mid vowels become closing diphthongs in open syllables, including [ɤɯ] for /ɤ/.

I'm also fairly sure some language had [ɑː] > [ɑu], and I know Mandarin had [ɑ] > [uɔ] in some environments. Basically what I'm saying is diphthongs arising from [ɑ] are likely to have a rounded component since many languages sort of treat open vowels as unspecified for rounding. Alternatively, due to how readily open vowels shift into other open vowels, any outcome from [a] is plausible.

Zaarin wrote:I have a rather weird one I want to ask about. s/k/_# Does that seem too far fetched?

maybe, although I have heard that some natlangs have some pretty strange sound changes too

However, it seems that s > t / _# has happened in some natlangs, and t > k is not unheard of too.

Another way:

step 1: s > h / _#
step 2: h > q / _#
step 3: q > k

one can also read the Index Diachronica for answers: https://chridd.nfshost.com/diachronica/

Frislander wrote:

Frislander wrote:I'm back-deriving the proto-inventory for this language, and I have a phoneme /ʈʳ/ and I want it to split and caused a position-based alternation in the daughter language. Does changing it to /t/ word-initially and /t͡ʃ/ otherwise seem sensible?

I think this was buried, but my question still stands: does this seem OK?

Additional information: this change comes in the context of a group of changes related to fricatives which sees /ɸ/ become /p/ initially and /w/ otherwise, /ʃ/ becoming /s/ initially and /j/ otherwise, and /ʂ/ to /s/ initially and /ɽ/ otherwise.

What happens to the other stops? /p t k/ or whatever you have?

ʈʳ > t/#_ is in line with your other changes in the retroflex POA so this is non-controversial. You're probably more worried about ʈʳ > t͡ʃ am I right? A problem I see with that is that you have ʂ > ɽ in the same position, whereas you want to have a movement in the opposite direction with the stop. Do you have ɽ in the proto-inventory? What happens to it?

gestaltist wrote:

Frislander wrote:

Frislander wrote:I'm back-deriving the proto-inventory for this language, and I have a phoneme /ʈʳ/ and I want it to split and caused a position-based alternation in the daughter language. Does changing it to /t/ word-initially and /t͡ʃ/ otherwise seem sensible?

I think this was buried, but my question still stands: does this seem OK?

Additional information: this change comes in the context of a group of changes related to fricatives which sees /ɸ/ become /p/ initially and /w/ otherwise, /ʃ/ becoming /s/ initially and /j/ otherwise, and /ʂ/ to /s/ initially and /ɽ/ otherwise.

What happens to the other stops? /p t k/ or whatever you have?

ʈʳ > t/#_ is in line with your other changes in the retroflex POA so this is non-controversial. You're probably more worried about ʈʳ > t͡ʃ am I right? A problem I see with that is that you have ʂ > ɽ in the same position, whereas you want to have a movement in the opposite direction with the stop. Do you have ɽ in the proto-inventory? What happens to it?

The other stops (/p t k/) are unaffected by all changes bar a set of palatalisations before /i/, most of which merge with that intervocalic ʈʳ (with the latter probably through a medial /ʈʂ. stage) to form a new set of palatal affricates not present in the protolang. My problem was actually more with the change ʈʳ > t/#_, but if you're fine with it then I'm OK. /ɽ/ is not present in the proto-language, but it had /l/, which then became /ɽ/, thus merging with intervocalic /ʂ/.

Frislander wrote: The other stops (/p t k/) are unaffected by all changes bar a set of palatalisations before /i/, most of which merge with that intervocalic ʈʳ (with the latter probably through a medial /ʈʂ. stage) to form a new set of palatal affricates not present in the protolang. My problem was actually more with the change ʈʳ > t/#_, but if you're fine with it then I'm OK. /ɽ/ is not present in the proto-language, but it had /l/, which then became /ɽ/, thus merging with intervocalic /ʂ/.

Sounds good, then.

What are known ways for a language that doesn't already have uvulars to develop a uvular stop (and maybe also fricative) series? I know some Turkic languages have a palatal:velar alternation in their dorsal stops (largely related to neighboring vowels) while others have velar:uvular, and I don't know which is older; did some Turkic languages back original velars to uvulars in some environments, or were the uvulars there from the beginning with the palatal:velar alternation being a case of both velars and uvulars fronting?

(This post brought to you by my undying love of uvular consonants)

Chengjiang wrote:What are known ways for a language that doesn't already have uvulars to develop a uvular stop (and maybe also fricative) series? I know some Turkic languages have a palatal:velar alternation in their dorsal stops (largely related to neighboring vowels) while others have velar:uvular, and I don't know which is older; did some Turkic languages back original velars to uvulars in some environments, or were the uvulars there from the beginning with the palatal:velar alternation being a case of both velars and uvulars fronting?

(This post brought to you by my undying love of uvular consonants)

One possible way is k > q, t,c > k. Arabic and Aramaic got /q/ from /kˁ/, which was, of course, historically /kʼ/. k > q before low vowels has also been known to happen; have a vowel shift and you now have phonemic uvulars. x ɣ > χ ʁ is well attested. And, of course, r > ʁ~ʀ happened in Western Europe; ʁ > q is attested in some forms of Inuit.

Most uvulars I know of come from velars near back vowels, or occasionally contact with pharyngeals or other uvulars - e.g. French /kr/ showing up sometimes as [qX]. Given how commonly dorsal frivatives vary between velar and uvular, that might be a source. Armenian shifted its dark lateral to a uvular frivative, and I've got something very similar for some coda /l/ as well. And they already had uvulars, but Caucasian languages often have lateral frivatives/affricates that shifted to velars or uvulars, especially in the Lezgian branch I believe.

I'm interested in whether a certain series of vowel changes seems like total baloney or not. Basically the way it starts is a > aə in open syllables or something, and then we get:

aə > ɑɨ > ɔi

Basically, the first element of the diphthong backs while the second element raises and fronts. Does this make sense? I'm a lot less well-versed in vowel changes than I am with consonant changes.

It makes sense to me.

Do you know of any precedent for the change(s)?

I would imagine that what happens is that /aə/ is analysed as a sequence of vowels and /a/ becomes a back vowel while /ə/ goes through its own series of changes. Nearby consonants can make allophonic changes that are then generalised as an alternative.

thetha wrote:Do you know of any precedent for the change(s)?

I don't really have historical sound changes memorised and I don't know much out of Romance and Germanic anyway. I recommend searching the Index Diachronica.

There's */raʔ/ > /lɵy/ from Old Chinese to Cantonese, in the same general direction, even though not exactly the same.

Don't some dialects of English (Australian? can't remember exactly...) have /ai au/ [ɑɪ æʊ] or similar? From there on it's a rather small step to something like [ɔɪ ɛʊ].

Cedh wrote:Don't some dialects of English (Australian? can't remember exactly...) have /ai au/ [ɑɪ æʊ] or similar?

Yes, apparently Australian has [ɑe] and [æɔ]. My American English has roughly the realizations you mentioned FWIW though I don't know how mainstream that is. [ɔɪ] for /ai/ is also present in some accents.

So I recently found out that some danish dialects have gone through this lovely set of changes:

iː yː uː > iˀ yˀ uˀ (the developement of "stød") More vowels than just these and in almost all dialects including standard Danish. Occured via a pitch accent system, had a rather complex conditioning environment, sometimes incuded short vowels and only in Old Danish monosyllables + certain loanwords.

iˀ yˀ uˀ > ikʲ(ə) ykʲ(ə) ukʷ(ə) ("klusilspringet") in the dark green areas of this map http://jyskordbog.dk/ordbog/scripts/atl ... cgi?nr=2_1 . In other dialects ix(ə) yx(ə) ux(ə) or itʲ(ə) ytʲ(ə) uk(ə), sometimes with a conditioning environment of _# or _[#s]. Some dialects (Vest-hardsyssel and Fanø) also eˀ oˀ > ejkʲ(ə) owkʷ(ə) ((or maybe ejg̥ʲ(ə) owg̥ʷ(ə), source is unclear) (øˀ > øjkʲ(ə) only sporadically)

ikʲ(ə) ykʲ(ə) ukʷ(ə) > iː yː uː Quite recent and not in all dialects. EDIT: in some dialects to iˀ yˀ uˀ.


How realistic do you think it would be to have the first change occur in more environments and also in polysyllabic words but only on long vowels and have the third change not occur at all? What if the vowel system is still big but less crazy than the Danish one?

Still trying to work out my vocalic system.
The proto-language has a very turkish-like:
i y ɯ u
ɛ ø a o
All of these can be long or short, vowel length is phonemic in the proto-lang. There is vowel harmony.
Possible dipthongs are Vj, Vw, other things which might condition the vowels are /Vɣ/, /Vh/, /Vɣʷ/, /VCʲ/
/j/ and /w/ are transparent to vowel harmony in the language and can appear in both [+front] and [-front] words.
I'm looking to have some vowel shift that changes around this system somewhat, destabilize it so it shifts to some other kind of harmony or so the qualities change, but nearly all the resources I can find on vowel shifts are for Germanic languages.

Here is a collection of information about vowel shifts.

Is this fine ?

K = velar
Kʷ = labialized velar

uK oK > ɯKʷ ɤKʷ

ɯ ɤ > ə