Sound Change Quickie Thread

[Zaarin]

I can't cite you a source, but I'm almost certain some language (I want to say it's either Afro-Asiatic or Native American, which I realize isn't very helpful but those are the two groups I've studied recently ) has j > w word-initially.

I think I read it either on the board or on Wikipedia, but Hebrew is thought to have had the opposite happen: *w- > *j-.

That's probably what I'm thinking of then.

[Bristel]

What's something wacky I can do with /j/ that isn't just fortition to /dZ/?

You could have it disappear word-initially.

[hwhatting]

What's something wacky I can do with /j/ that isn't just fortition to /dZ/?

You could have it disappear word-initially.

Yes, or turn it into /h/, as Greek did in the cases where it didn't fortify it to /d_z/.

[Dē Graut Bʉr]

What can clicks do?

[Birdlang]

What can the sje-sound turn into? I am working on a language family and the protolang has that phoneme. The sje-sound is ɧ and its romanization is a j with stroke from Unicode.
as well as ingressive fricatives (s and z) and what can uvularized sounds turn into? I have uvularized b, y, z, v, f, s, h

[Dē Graut Bʉr]

Some immigrants pronounce the sje-sound as /χ/, so that'd be a possibility.

[Porphyrogenitos]

What sound changes are made more or less likely by a vowel being in a closed or open syllable? For example, I know vowel breaking is more common in open syllables, at least among the Romance languages. Are there other vowel changes, mergers, etc, that are affected by syllables' closed/open-ness?

[vokzhen]

What sound changes are made more or less likely by a vowel being in a closed or open syllable? For example, I know vowel breaking is more common in open syllables, at least among the Romance languages. Are there other vowel changes, mergers, etc, that are affected by syllables' closed/open-ness?

That qualifier is important, as I don't believe it holds up as a rule outside of Romance. Irresponsibly taking the wikipedia page as representative gives the opposite idea, most of the examples they give are closed syllables. In reality I'm not sure there's a tendency one way or the other.

Open syllables can have spontaneous lengthening, and closed shortening (which can create new phonemes, e.g. /i ii/ [ɪ i:] > [ɪ i i:] as a result of shortening). Somewhat contra to that, I get the impression that medial closed syllables are less likely to reduce to zero than medial open ones, probably because of the CCC clusters it would create.

If the language is structured a particular way (monosyllabic, CV(C) structure), tonogenisis from losing certain coda consonants might end up looking kind of like something that happened to open or closed syllables. Lose coda /h ?/, open syllables have three tones (low from h, middle from original open syllables, high from ?) while closed have a single fixed tone.

[Qwynegold]

I get too many ideas lately. Now I want to turn Swedish into a Northwest Caucasianesque language. There are two processes in this language that I want to ask about. One is vowels dumping their features on preceding consonants, and one is something similar to stød in Danish, except here the pitch accent turns into pharyngealization of consonants.

• What should I do about labiovelarized and labiopalatalized labials, which I find awkward? They are /mʷ, pʷ, bʷ, fʷ, vʷ, mᶣ, pᶣ, bᶣ, fᶣ, vᶣ + pharyngealized variants of all these/. I could do /fʷ, vʷ, fᶣ, vᶣ/ /ɸ, β, ɸʲ, βʲ/, but what about the rest? Or should I just leave them all as is? Or just drop the rounding?
• What should I do about consonants that are simultaneously in a palatalizing environment and a pharyngealizing environment? Would these two things just cancel eachother out, or what do you think should happen?

Aw, anyone?

[Herr Dunkel]

As for 1) you could either drop the rounding or make it into an offglide like [pʷa pwa] and then you can do all sorts of stupid stuff with it and 2) I think yeah it could cancel out

[Birdlang]

Thank you. I will turn the uvularized consonants into uvular fricative released consonants and the sje sound into X.

[vokzhen]

• What should I do about labiovelarized and labiopalatalized labials, which I find awkward? They are /mʷ, pʷ, bʷ, fʷ, vʷ, mᶣ, pᶣ, bᶣ, fᶣ, vᶣ + pharyngealized variants of all these/. I could do /fʷ, vʷ, fᶣ, vᶣ/ /ɸ, β, ɸʲ, βʲ/, but what about the rest? Or should I just leave them all as is? Or just drop the rounding?
• What should I do about consonants that are simultaneously in a palatalizing environment and a pharyngealizing environment? Would these two things just cancel eachother out, or what do you think should happen?

Aw, anyone?

I find it useful to split NWC-type languages into the consonant features and the consonant-plus-vowel features (i.e. /l/ has no secondary features on its own, but the vowel after it is always in... one of those languages).

For labials, then, one option would be to not have a plain series, or have the plain series round the following vowel, depending on how you want to look at it. The fronted set could stay distinct by not requiring rounded. As a result, you have /mɨ mʲɨ mᶣɨ/ as [mu mi my] without the option for just [mɨ].

For the palatalizing+pharyngealizing, I have a feeling they would collapse, but not straightforwardly. As a few examples:
t > /t/ [tɨ]
tʲ > /ts/ [tsi]
tᶣ > /tsʷ/ [tsy]
tˁ > /tˁ/ [tˁɯ]
tˁʲ > /tsˁ/ [tsˁɛ]
tˁᶣ > /tsʷˁ/ [tsˁœ]
k > /k/ [kɨ]
kʲ > /tɕ/ [tɕi]
kᶣ > /tɕʷ/ [tɕʷy]
kˁ > /kˁ/ [kˁɯ]
kˁʲ > /tɕˁ/ [tɕˁɛ]
kˁᶣ > /tɕʷˁ/ [tɕʷœ]
The /ts/ and /tɕ/ series always take the front allophones of vowels in this scheme, vowel quality remains allophonic. Instead of a highly symmetric /t tʲ tᶣ tˁ tˁʲ tˁᶣ k kʲ kᶣ kˁ kˁʲ kˁᶣ/, you have a more asymmetric /t tˁ ts tsʷ tsˁ tsˁʷ tɕ tɕʷ tɕˁ tɕˁʷ k kˁ/, while avoiding (straightfoward) palatalized+pharyngealized segments. Which when you look at NWC really looks like what happened, as there's "missing" consonants all over the coronal area. However, given that Abkhaz has [ɥˁ] for Abaza /ʕʷ/, I think you're also safe overlapping pharyngealization and palatalization if you want to keep it more symmetrical and straightfoward. Another option would be to accept "pharyngealization" as shorthand for consonant-driven vowel allophony, even if the consonants themselves don't exhibit pharyngeal constriction. If feature-dropping preceded pharyngealization, though, I wouldn't dismiss pharyngealization staying on the vowel, so instead of /ɨ ə a/ + a whole set of pharyngealized consonants you have /ɨ ɨˁ ə əˁ a aˁ/.

[Herr Dunkel]

I'm thinking of doing a-umlaut in the form of /ø ø̃ ø: ø̃:/ [ɶ ɶ̃ ɶ: ɶ̃:] / _${a ã}, any precedents outside the old {i u} {e o} / _${a ã}?

[Qwynegold]

As for 1) you could either drop the rounding or make it into an offglide like [pʷa pwa] and then you can do all sorts of stupid stuff with it and 2) I think yeah it could cancel out

Hmm, what kind of stuff?

I find it useful to split NWC-type languages into the consonant features and the consonant-plus-vowel features (i.e. /l/ has no secondary features on its own, but the vowel after it is always in... one of those languages).

For labials, then, one option would be to not have a plain series, or have the plain series round the following vowel, depending on how you want to look at it. The fronted set could stay distinct by not requiring rounded. As a result, you have /mɨ mʲɨ mᶣɨ/ as [mu mi my] without the option for just [mɨ].

For the palatalizing+pharyngealizing, I have a feeling they would collapse, but not straightforwardly. As a few examples:
t > /t/ [tɨ]
tʲ > /ts/ [tsi]
tᶣ > /tsʷ/ [tsy]
tˁ > /tˁ/ [tˁɯ]
tˁʲ > /tsˁ/ [tsˁɛ]
tˁᶣ > /tsʷˁ/ [tsˁœ]
k > /k/ [kɨ]
kʲ > /tɕ/ [tɕi]
kᶣ > /tɕʷ/ [tɕʷy]
kˁ > /kˁ/ [kˁɯ]
kˁʲ > /tɕˁ/ [tɕˁɛ]
kˁᶣ > /tɕʷˁ/ [tɕʷœ]
The /ts/ and /tɕ/ series always take the front allophones of vowels in this scheme, vowel quality remains allophonic. Instead of a highly symmetric /t tʲ tᶣ tˁ tˁʲ tˁᶣ k kʲ kᶣ kˁ kˁʲ kˁᶣ/, you have a more asymmetric /t tˁ ts tsʷ tsˁ tsˁʷ tɕ tɕʷ tɕˁ tɕˁʷ k kˁ/, while avoiding (straightfoward) palatalized+pharyngealized segments. Which when you look at NWC really looks like what happened, as there's "missing" consonants all over the coronal area. However, given that Abkhaz has [ɥˁ] for Abaza /ʕʷ/, I think you're also safe overlapping pharyngealization and palatalization if you want to keep it more symmetrical and straightfoward. Another option would be to accept "pharyngealization" as shorthand for consonant-driven vowel allophony, even if the consonants themselves don't exhibit pharyngeal constriction. If feature-dropping preceded pharyngealization, though, I wouldn't dismiss pharyngealization staying on the vowel, so instead of /ɨ ə a/ + a whole set of pharyngealized consonants you have /ɨ ɨˁ ə əˁ a aˁ/.

Hmm... But I wan't to keep palatalization. But you've given me some ideas, thanks! If there's precedence for [ɥˁ], then maybe e.g. tɕˁ isn't so bad.

[Qwynegold]

What can the sje-sound turn into? I am working on a language family and the protolang has that phoneme. The sje-sound is ɧ and its romanization is a j with stroke from Unicode.
as well as ingressive fricatives (s and z) and what can uvularized sounds turn into? I have uvularized b, y, z, v, f, s, h

Oh, I was gonna respond to this but forgot. There aren't any exactly any natlang examples of it turning to anything, but I think S and x are most likely. f or W might also happen.

[Hallow XIII]

What can the sje-sound turn into? I am working on a language family and the protolang has that phoneme. The sje-sound is ɧ and its romanization is a j with stroke from Unicode.
as well as ingressive fricatives (s and z) and what can uvularized sounds turn into? I have uvularized b, y, z, v, f, s, h

Oh, I was gonna respond to this but forgot. There aren't any exactly any natlang examples of it turning to anything, but I think S and x are most likely. f or W might also happen.

Aren't there? What about those Swedish dialects that have X_w for <sj> (or is that conservative)?

[احمکي ارش-ھجن]

How to turn this:
/p t tʲ k kʲ q ʔ/<p t ṭ k ḳ q ʾ>
/b d dʲ g gʲ/<b d ḍ g ġ>
/m n/<m n>
/f s sʲ x xʲ/< f s ṣ h ḥ>
/w r l j/<w r l j>
/ɬ/<ḻ>

Into this:
/m n ŋ ŋʷ/<m n ng nw>
/p p’ b t t’ d k k’ g kʷ k’ʷ gʷ q q’ ʔ/<p ṗ b t ṭ d k ḳ g kw ḳw gw c ċ q>
/f v s z x ɣ xʷ~ʍ ɣʷ~w/<f v s z x j wh w>
/t͡s t͡s’ k͡x k͡x’/<tz ṭz kg ḳg>
/l r j/<l r y>

In other words: how to get rid of palatalized consonants and then how to develop ejectives and labialized velars?

[Hallow XIII]

that question is hilariously malformed

[احمکي ارش-ھجن]

Malformed...

[Hallow XIII]

By which I mean, you don't give anywhere near enough information about the structure of words in your language for anybody to be able to give you good and useful answers. Lists of consonants really aren't enough material for that; you'd need at least a description of the vowels, syllable structure, stress patterns, maybe common morphemes and constructions and so on...

And I suspect, quite frankly, that the answer would be "you don't, at least not in any reasonable time".

[احمکي ارش-ھجن]

Well, I have vowels:
/a e i o u/<a e i o u>
/a: e: i: o: u:/<ā ē ī ō ū>
/á é í ó ú/<á é í ó ú> <: pitch accent
They are the same as they are in the proto-language

And the syllable structure is (C)(C)V(C)(C), though CV, CVC, and CCV are most common.

Also stress is penultimate

[Pogostick Man]

In other words: how to get rid of palatalized consonants and then how to develop ejectives and labialized velars?

You can get labialized velars from sequences of velar + vowel (preferably rounded IIRC) + glide/laryngeal. There's precedent for this in the Mayan languages, I think.

[vokzhen]

Yea, kuj > kʷi and so on. Depending on how common they're meant to be, you could do something like u > i and/or o > e, with original ku/o becoming kʷu/o > kʷi/e. The u > i is the more likely of the two to be unconditioned, o > e I'd think would probably have to be via i-mutation or fronting before coronals, or maybe o > ou > eu (dissimilation) > e(i). Total syllable loss might be a possibility as well.

For ejectives, coalescence with ʔ is the least complicated (though if you go that route, more than likely - at least at first - you've have sequences like k+ʔ>[k'(ʔ)] and m+ʔ>[m̰(ʔ)], so it's interpretation that the latter is a cluster and the former isn't). You can also do it via debuccalization of stop-stop sequences to make new glottal stops. If high pitch can have laryngealization in the proto-language that could shift to consonants in one branch and be lost in another, but again, I'd expect to have glottalized sonorants as well. If the voiceless consonants are aspirated in some or all positions, there are potential complications with that (i.e. a number of languages aspirate stops before other stops and word-finally), but if there is some kind of opposition rather than across-the-board aspiration, coming in close contact with a language with an aspirate-ejective-voiced system might cause the plain allophones to become ejective and phonemicize the aspirated ones.

On a different note, whyyyyy would you use dots for both palatalization and ejectivization

[احمکي ارش-ھجن]

On a different note, whyyyyy would you use dots for both palatalization and ejectivization

Because they are two different languages with two different romanizations.

I suppose high pitch can cause laryngealization, whereas in my other language Vrkhazhian, low tone caused creaky voice, which caused uvularization/pharyngealization of preceding consonants.

There are two modern languages in my story, Ngu-Cam, which is the language in question, and Vrkhazhian, a distant cousin. Both of them are descendants of Proto-Hasjakam, though they split up into two branches after the Himoshian language stage.
In other words:
1) Hasjakam Himoshian Highalic Proto-Vrkhazhian Old Vrkhazhian Middle Vrkhazhian Modern Vrkhazhian
2) Hasjakam Himoshian Nan-Tmar ??? ??? Ngu-Cam
But then that would mean they'd have to share the same sound change from Hasjakam to Himoshian:
1a) Low tone > creaky voice > pharyngealised preceding consonant
1b) /tʲ dʲ sʲ kʲ gʲ xʲ/ > /ts dz ʃ c ɟ ç/

[jmcd]

אקֿמך ארש-הגִנו: Which consonant clusters are most common? Are all consonant clusters found in all positions? Are there consonant culsters which appear only in loanwords? And then there's the possbility of interaction between grammatical change and phonological change.