Sound Change Quickie Thread

[Dē Graut Bʉr]

Yeah, I want to have a sound change that didn't make sense. It took something like 5000 or 6000 years, so quite a lot can have happened before it finally became a k’, I guess.

[Drydic]

Yeah, I want to have a sound change that didn't make sense. It took something like 5000 or 6000 years, so quite a lot can have happened before it finally became a k’, I guess.

Well fuck, gimme 5-6000 years and I can turn Chinese into Navajo and go through Arabic along the way...

[CatDoom]

Could anyone explain how /z/ possibly can have become /k’/ in one of my conlangs?

Well, /t/ merged unconditionally with /k/ in Hawaiian in far less than 5000 years, and that's a pretty big shift in place of articulation. Malagasy, another Austronesian language, picked up a Sandhi-type alternation between fricatives and stops in certain positions, including /h/ alternating with /k/. That change presumably happened over the course of less than 2,000 years, and possibly less than 1,000.

So lets say that at an early stage /z/ undergoes debuccalization to /ɦ/. Phonemic distinctions between /ɦ/ and /h/ are pretty rare, as far as I can tell, so it doesn't seem unreasonable that /ɦ/ would wind up being realized as /h/. Subsequent to that the language could have developed an alternation between /h/ and /k/ as in Malagasy.

I'm honestly not sure how ejective consonants enter a language's phonemic inventory, but Yapese (another Austronesian language) gained them somehow. Maybe /k/ as an allophone of the glottal /h/ could become the glottalized consonant /k'/ through dissimilation, making it contrast with a /k/ phoneme occurring in other positions. At that point the alternation might break down over time, causing /h/ to merge to /k'/ in all positions.

Of course, that series of sound changes seems like it would probably result in other fricatives turning into /k'/ as well... I'm not sure why /z/ would be debuccalized if /s/ wasn't. I suppose it's possible that, at least at the time, the language didn't actually have /s/ as a phoneme, which wouldn't necessarily prevent it from gaining one later. Sound changes have no memory, after all, and there are a number of ways that /h/ could have re-entered the language's consonant inventory as well, after it had previously merged to /k'/.

[Nortaneous]

I'm honestly not sure how ejective consonants enter a language's phonemic inventory, but Yapese (another Austronesian language) gained them somehow.

*CVC syllables with q (which was probably a uvular stop in Proto-Austronesian but turned to a glottal stop in most languages, presumably including Yapese) as the second consonant collapsed to *CC clusters and then to C’:
*mataqu > mat’aaw
*taqi > t’aay

Yapese ejective fricatives are usually analyzed as F + glottal stop sequences.

[8Deer]

Could anyone explain how /z/ possibly can have become /k’/ in one of my conlangs?

The only thing I can imagine is something like z > ʒ (or maybe dz) > dʒ > ɟ > g, and then somehow get from g > k'. Maybe your voiced stops are phonetically implosive, and then shift to ejective? But this whole shift involves some fairly implausible steps so I have my doubts.

[Nortaneous]

z > r > gʟ > g > ɠ > k’

[Nesescosac]

I don't know what we'd do if it were not for Hiw.

[Pogostick Man]

I don't know what we'd do if it were not for Hiw.

Were't not for Hiw, what would we do?
We'd bash the phonemes with a shoe
So we could make our phonemes do
The sound changes we'd want them to
Through sequences impossibru;
We'd end up with phonetic stew
And tear all of our hair out too
But now we have an out, for—phew!—
Nortaneous has heard of Hiw.

[Drydic]

Could anyone explain how /z/ possibly can have become /k’/ in one of my conlangs?

The only thing I can imagine is something like z > ʒ (or maybe dz) > dʒ > ɟ > g, and then somehow get from g > k'. Maybe your voiced stops are phonetically implosive, and then shift to ejective? But this whole shift involves some fairly implausible steps so I have my doubts.

*dw > erk

Nothing is true, everything is permitted.

[Click]

Could anyone explain how /z/ possibly can have become /k’/ in one of my conlangs?

z → ð ~ r → ɣ → ɡ → kʼ

[Dē Graut Bʉr]

Thank you for your answers! As all of them go through *g it seems I have to devoice the voiced plosives of the proto-language, and aspirate the old voiceless plosives. And while I was working on it anyway, I've also added some new sound changes. It becomes more and more difficult to believe that it in fact is derived from an IE-ish proto-language!

[R.Rusanov]

z > ts > t' > k'

[Tropylium]

A route designed to get as little as possible in the way of anything else:

z → ɮ → d͡ɮ → dl → dɫ → dɣ → ʔɣ → ʔx → xʼ → kʼ

[Click]

I'm slowly working on diachronics for my Littoran family and I've run into some problems., so is a ʃ ʒ → θ ð shift plausible?

[sangi39]

I'm slowly working on diachronics for my Littoran family and I've run into some problems., so is a ʃ ʒ → θ ð shift plausible?

I suppose if you went through some kind of apical /s z/ vs. laminal /s z/ stage in between then you could end up with /θ ð/ without to many problems (didn't Spanish /θ/ come from laminal or apical /s/? I forget which)

[WeepingElf]

I'm slowly working on diachronics for my Littoran family and I've run into some problems., so is a ʃ ʒ → θ ð shift plausible?

I suppose if you went through some kind of apical /s z/ vs. laminal /s z/ stage in between then you could end up with /θ ð/ without to many problems (didn't Spanish /θ/ come from laminal or apical /s/? I forget which)

Spanish /θ/ is from laminal /s/, I think, but I am not sure.

[sangi39]

I'm slowly working on diachronics for my Littoran family and I've run into some problems., so is a ʃ ʒ → θ ð shift plausible?

I suppose if you went through some kind of apical /s z/ vs. laminal /s z/ stage in between then you could end up with /θ ð/ without to many problems (didn't Spanish /θ/ come from laminal or apical /s/? I forget which)

Spanish /θ/ is from laminal /s/, I think, but I am not sure.

Wikipedia would seem to agree with you. In that case you could have something like:

ʃ ʒ > s̪ z̪ > θ ð

While /s̺ z̺/ remain throughout the period.

[Click]

I'm slowly working on diachronics for my Littoran family and I've run into some problems., so is a ʃ ʒ → θ ð shift plausible?

I suppose if you went through some kind of apical /s z/ vs. laminal /s z/ stage in between then you could end up with /θ ð/ without to many problems (didn't Spanish /θ/ come from laminal or apical /s/? I forget which)

Spanish /θ/ is from laminal /s/, I think, but I am not sure.

Wikipedia would seem to agree with you. In that case you could have something like:

ʃ ʒ > s̪ z̪ > θ ð

While /s̺ z̺/ remain throughout the period.

Thanks!

[Richard W]

what can you do with implosives? some ideas i've had are:
/ɓ ɗ ɠ/ → /mb nd ŋɡ/
/ɓ ɗ ɠ/ → /b d ɡ/
are these plausible?

Proto-Tai preglottalised stops /ʔb ʔd/ seem to go (merge) straight to /m n/ without a prenasalised stage. /ʔd/ can also become /l/; Shan has /ʔb ʔd/ > /m l/. Proto-Tai lacks /ʔɡ/. /ʔb/ > /w/ is also known from Shan (details unknown to me- there’s a lot of variation across Shan).

Preglottalised stops (and thus, I presume, implosives) can remain a reservoir of voiced stops if you choose to devoice your voiced stops.

[Basilius]

Proto-Tai preglottalised stops /ʔb ʔd/ seem to go (merge) straight to /m n/ without a prenasalised stage. /ʔd/ can also become /l/; Shan has /ʔb ʔd/ > /m l/. Proto-Tai lacks /ʔɡ/. /ʔb/ > /w/ is also known from Shan (details unknown to me- there’s a lot of variation across Shan).

However, the real question is: why is that series reconstructed as preglottalized rather than prenasalized or something else (e. g. nasal with some non-default phonation)?

[Richard W]

Proto-Tai preglottalised stops /ʔb ʔd/ seem to go (merge) straight to /m n/ without a prenasalised stage. /ʔd/ can also become /l/; Shan has /ʔb ʔd/ > /m l/. Proto-Tai lacks /ʔɡ/. /ʔb/ > /w/ is also known from Shan (details unknown to me- there’s a lot of variation across Shan).

However, the real question is: why is that series reconstructed as preglottalized rather than prenasalized or something else (e. g. nasal with some non-default phonation)?

The reasons I find it persuasive are:

• *ʔb and *ʔd pattern with *ʔ (> /ʔ/) and *ʔj (> /j/ or /ʔj/) when it comes to their effect on how the tones split.
• [ʔb] and [ʔd] remain in a few dialects, e.g. Wuminɡ in Northern Zhuang. (Thai orthography showed the glottal stop in its reflex of *ʔj, though now only in four words.)
• Sui /hm/ and /ʔm/ correspond to Proto-Tai *m̥.
• Proto-Tai *ʔj and *ɲ either retain their contrast (e.g. Lao) or both merge with *j (e.g. Thai), arguing against pre-nasalisation.

[sangi39]

Are the following sound changes plausible, realistic, etc.

*p *t *tʃ *k *kʷ > /pʰ tʰ tʃʰ kʰ kʷʰ/ except after /N/
*mp *nt *ntʃ *ŋk *ŋkʷ > /mb nd ndʒ ŋg ŋgʷ/
*b *d *dʒ *g *gʷ > /p t tʃ k kʲ kʷ/ except after /N/
*mb *nd *ndʒ *ŋg *ŋgʷ > /b d dʒ g gʷ/

preceding /r l/ prevent the voicing or aspiration of plosives, e.g. *rp *lt *rg > /rp lt rg/

[Nortaneous]

mb > b > p > ph chain shift, sure, followed by mp > mb voice assimilation to fill the gap

dunno about aspiration blocking but stranger things have happened

[sangi39]

mb > b > p > ph chain shift, sure, followed by mp > mb voice assimilation to fill the gap

I didn't even notice the chain shift until you pointed it out. I guess the formatting kind of made me miss it

dunno about aspiration blocking but stranger things have happened

Hopefully someone might know a bit more, but I'm keeping it for the moment

[Nortaneous]

yeah i'd say keep it, call it spreading of some sort of lenis feature. voiced stops not devoicing after voiced consonants is completely fine, and given that plosives didn't aspirate after s in germanic, something vaguely similar should work.