Sound Change Quickie Thread

[Cael]

i was wondering. How would you all derive palatal phonemes from non-palatalization?

[WeepingElf]

also, I'd appreciate it very much if anyone would be so kind as to suggest a consonant that can be plausibly turned into /k/ and /l/.

In my Hesperic family, I have "Drummond's Law" (named after a fictional linguist), according to which a consonant followed by a laryngeal is velarized (and the laryngeal is lost). So you get a velarized /ł/ out of /l/+laryngeal. In the Albic branch, this /ł/ has become /ɣ/ and later /g/, which has become /k/ in one dialect of Old Albic. (Non-velarized /l/ still is just /l/ in Albic.)

You can easily have /ł/ become /k/ by a sequence of changes like the above in some contexts (or some daughter languages), and plain /l/ in others. I don't know a natlang precedent for the whole sequence, but I am pretty sure that there are natlang precedents for every single sound change in the sequence.

EDIT: IPAfied the phonemes discussed.

[Qwynegold]

Would ħ > x be plausible with a language that also has χ and h?
also, I'd appreciate it very much if anyone would be so kind as to suggest a consonant that can be plausibly turned into /k/ and /l/.

I don't know about it skipping over χ like that. I would rather expect:
χ > x
ħ > χ

/x̞ʷ/ > /ɰʷ/ > /w/

Does that seem about right?

ɰʷ is basically the same thing as w!

[Nortaneous]

Would ħ > x be plausible with a language that also has χ and h?
also, I'd appreciate it very much if anyone would be so kind as to suggest a consonant that can be plausibly turned into /k/ and /l/.

I don't know about it skipping over χ like that. I would rather expect:
χ > x
ħ > χ

nah, chalk it up to Weird Shit. abkhaz had ʕʷ > ɥ, so.

or, say your χ is harder, like [ʀ̥] or so, then it'd be reasonable for [ħ], a 'softer' sound, to skip it and jump to [x].

[TaylorS]

in the evolution of the new Mitian (thanks for inventing the term, WeepingElf!) diachronic lang Karpasian that I am working on there is a sound change between vowels:

/b d ʥ g/, / f s ɕ h/ > /v z ʑ Ø/
/p t ʨ k/, /b: d: ʥː g:/ > /b d ʥ g/
/p: t: ʨː k: f: s: ɕː x:/ > /p t ʨ k f s ɕ h/
/l/ > /r/

This shift results in Finnish-stype consonant gradiation in the morphophonology.

so abbanuter "of the mothers" becomes abanuder.

[Whimemsz]

i was wondering. How would you all derive palatal phonemes from non-palatalization?

Palatalization in proximity to some sort of front vowel or glide (/i/, /e/, /j/, etc.) is extremely common; you can make it phonemic through vowel loss, vowel shifts/splits, a large number of nativized loanwords, or the loss of /j/ in clusters etc. In Spanish, for example, velars were palatalized before front vowels and /j/ (these palatalized phones later became affricates [and then > fricatives] but the principle is the same), coronals were palatalized before /j/ which was then lost, the initial member of some clusters (such as -kt-) became /j/, which palatalized the following stop (and changed the preceding vowel) before disappearing, clusters of stop+l became palatals (though they again underwent further changes), and long -nn- and -ll- palatalized as well. So there you have five different instances of palatalization in just the history of Spanish (and with varying eventual outcomes).

[Cael]

i was wondering. How would you all derive palatal phonemes from non-palatalization?

Palatalization in proximity to some sort of front vowel or glide (/i/, /e/, /j/, etc.) is extremely common; you can make it phonemic through vowel loss, vowel shifts/splits, a large number of nativized loanwords, or the loss of /j/ in clusters etc. In Spanish, for example, velars were palatalized before front vowels and /j/ (these palatalized phones later became affricates [and then > fricatives] but the principle is the same), coronals were palatalized before /j/ which was then lost, the initial member of some clusters (such as -kt-) became /j/, which palatalized the following stop (and changed the preceding vowel) before disappearing, clusters of stop+l became palatals (though they again underwent further changes), and long -nn- and -ll- palatalized as well. So there you have five different instances of palatalization in just the history of Spanish (and with varying eventual outcomes).

Thanks Whimemsz. I didn't know palatalization could occur other than before /i/, /e/, /j/, etc. Now I see that it can originate from various environemnts. (P.S. I love your guide to polysynthesis. It really helped me plan out my conlang )

[suelior]

Would ħ > x be plausible with a language that also has χ and h?
also, I'd appreciate it very much if anyone would be so kind as to suggest a consonant that can be plausibly turned into /k/ and /l/.

I don't know about it skipping over χ like that. I would rather expect:
χ > x
ħ > χ

nah, chalk it up to Weird Shit. abkhaz had ʕʷ > ɥ, so.

or, say your χ is harder, like [ʀ̥] or so, then it'd be reasonable for [ħ], a 'softer' sound, to skip it and jump to [x].

This is awesome. I was actually going to have a plain-tense contrast and χ was the tense component, so this explanation should work out smoothly.

also, I'd appreciate it very much if anyone would be so kind as to suggest a consonant that can be plausibly turned into /k/ and /l/.

In my Hesperic family, I have "Drummond's Law" (named after a fictional linguist), according to which a consonant followed by a laryngeal is velarized (and the laryngeal is lost). So you get a velarized /ł/ out of /l/+laryngeal. In the Albic branch, this /ł/ has become /ɣ/ and later /g/, which has become /k/ in one dialect of Old Albic. (Non-velarized /l/ still is just /l/ in Albic.)

You can easily have /ł/ become /k/ by a sequence of changes like the above in some contexts (or some daughter languages), and plain /l/ in others. I don't know a natlang precedent for the whole sequence, but I am pretty sure that there are natlang precedents for every single sound change in the sequence.

EDIT: IPAfied the phonemes discussed.

Thank you very much for your input. Although, it would be nice if I could somehow avoid having to have /g/ > /k/.
Ah... I don't know, maybe I should stop trying to make one single all-encompassing root and according sound changes.

[Nortaneous]

Thank you very much for your input. Although, it would be nice if I could somehow avoid having to have /g/ > /k/.

Have /ɣ/ devoice.

[Tropylium]

A while ago, I invented a nice way to get a fairly bonkers compound soundlaw: voiced stops shift one POA forwards intervocally.
*b > β > w > gʷ > g
*d > ð > v > b
*ɖ > ɽ > ɾ > d
*ɟ > dʒ > dʐ > ɖ
*g > ɣ > j > ɟ
Better yet, this approach works for almost any stop inventory… retroflexes can be left out or replaced by postalveolars, labiovelars can easily be added, palatals can be left out if needed, etc.

---

As for a k ~ l correspondence, *tɬ in NE Caucasian yields just that, I believe.

[Nortaneous]

A while ago, I invented a nice way to get a fairly bonkers compound soundlaw: voiced stops shift one POA forwards intervocally.
*b > β > w > gʷ > g
*d > ð > v > b
*ɖ > ɽ > ɾ > d
*ɟ > dʒ > dʐ > ɖ
*g > ɣ > j > ɟ
Better yet, this approach works for almost any stop inventory… retroflexes can be left out or replaced by postalveolars, labiovelars can easily be added, palatals can be left out if needed, etc.

I might do something like this in a throwaway conlang... start with /p t c k b d ɟ g/ or so, have intervocalic lenition so the voiced stops fricate and the voiceless ones voice, then *w *ɣ > ɣ j and the intervocalic voiced stops fricate again, merging with what's already there. So...

b d ɟ g > w ð j ɣ / V_V
p t c k > b d ɟ g / V_V
j ɣ w > z j ɣ
ð > v (at this point, there'd probably be interference from a neighboring lang, since the coronal stops have to be alveolar for the next change)
b d ɟ g > v z j ɣ / V_V

So intervocalically, *p *d are reflected as /v/, *t *ɟ as /z/, *c *g as /j/, and *k *b as /ɣ/.

[WeepingElf]

A while ago, I invented a nice way to get a fairly bonkers compound soundlaw: voiced stops shift one POA forwards intervocally.
*b > β > w > gʷ > g
*d > ð > v > b
*ɖ > ɽ > ɾ > d
*ɟ > dʒ > dʐ > ɖ
*g > ɣ > j > ɟ
Better yet, this approach works for almost any stop inventory… retroflexes can be left out or replaced by postalveolars, labiovelars can easily be added, palatals can be left out if needed, etc.

I might do something like this in a throwaway conlang... start with /p t c k b d ɟ g/ or so, have intervocalic lenition so the voiced stops fricate and the voiceless ones voice, then *w *ɣ > ɣ j and the intervocalic voiced stops fricate again, merging with what's already there. So...

b d ɟ g > w ð j ɣ / V_V
p t c k > b d ɟ g / V_V
j ɣ w > z j ɣ
ð > v (at this point, there'd probably be interference from a neighboring lang, since the coronal stops have to be alveolar for the next change)
b d ɟ g > v z j ɣ / V_V

So intervocalically, *p *d are reflected as /v/, *t *ɟ as /z/, *c *g as /j/, and *k *b as /ɣ/.

Holy crap!

[communistplot]

/x̞ʷ/ > /ɰʷ/ > /w/

Does that seem about right?

ɰʷ is basically the same thing as w!

Thought as much, wasn't too sure then. But at least the /x_w/ --> /w/ is about right.

[finlay]

it pretty much happened directly like that in english - after all, 99% of us now pronounce 'wh' and 'w' the same.

[communistplot]

Ah, beautiful then, I do love some natlang precedent.

[Nortaneous]

Ah, beautiful then, I do love some natlang precedent.

better: x_w > h_w > w in some dialects of Welsh

[Thry]

Ah, beautiful then, I do love some natlang precedent.

better: x_w > h_w > w in some dialects of Welsh

And it's currently happening in my dialect of Spanish.

Juego or Juan /xwego/ and /xwan/ are realised as [hwego] and [hwaŋ] because /x/ has weakened to [h] everywhere. These, in turn, in some people's very casual speech, can be heard as [wego] and [waŋ] (or with [gw]). (which explains why some little kids being taught English will spell one as <juan> xD).

[Pinetree]

(which explains why some little kids being taught English will spell one as <juan> xD).

Which, to be fair, realy seems more logical than to spell it as if it were /'onɛ/

EDIT: If you were from that region, anyway.

[communistplot]

Ah, beautiful then, I do love some natlang precedent.

better: x_w > h_w > w in some dialects of Welsh

And it's currently happening in my dialect of Spanish.

Juego or Juan /xwego/ and /xwan/ are realised as [hwego] and [hwaŋ] because /x/ has weakened to [h] everywhere. These, in turn, in some people's very casual speech, can be heard as [wego] and [waŋ] (or with [gw]). (which explains why some little kids being taught English will spell one as <juan> xD).

Ah I see, I already have h_w become f unconditionally, though I guess I could condition it and have x_w merge with h_w and become w in other circumstances, maybe I could set up a chain shift:

h_w > f / #_ / C_
x_w > h_w
h_w > w

[Tropylium]

Ah I see, I already have h_w become f unconditionally, though I guess I could condition it and have x_w merge with h_w and become w in other circumstances, maybe I could set up a chain shift:

h_w > f / #_ / C_
x_w > h_w
h_w > w

Or you can just have hʷ > f occur before xʷ > hʷ does. That'd make it an actual chain shift.

(The word does not mean changes that apply to a single sound in succession; it means changes that apply to different sounds, so that each of the changes somehow triggers the next, either by producing something uncomfortably similar ("push chain"), or leaving a vacant place in the phonology ("pull chain").)

((Makes me wonder if we could call changes like *pʰ tʰ kʰ > ɸ θ x "swerve chains". In this case, one of the sounds changing should be able to trigger another doing the same due to the analogous locations in the sound system… and from what I've seen there's evidence that this is indeed how they proceed, not via a uniform instantaneous MOA switch. For example, Dutch appears to be stuck in the *β *ð > b d stage of Germanic initial voiced spirant fortition, and Vietnamese seems to have started on *pʰ > f but hasn't extended it to /tʰ/ yet.))

[communistplot]

Ah I see well I've settled on the changes then, and it'd fit well into what I was planning on doing anywho.

Initial Shift
f > h / #_ / _R
f > v / V_V
f > Ø / _#
f > [+geminate] / C_
h_w > f
x_w > h_w

Loss of h_w
h_w > hw
h > Ø / _(j, w) / _N

And I knew that, that's what I meant that the loss of /h_w/ would trigger the movement of /x_w/, the change afterwards is separate.

[Jetboy]

Is there any sort of precedent for [l] -> [r] and [ll] -> [l]?

[Pinetree]

Is there any evidence that allophones would become distinguished from the phoneme, and then merge with a different phoneme?

e.g. [ɾ] is an allophone of [d], then is considered distinct, then merges with [r]

[Thry]

Is there any sort of precedent for [l] -> [r] and [ll] -> [l]?

I think Romanian may have something similar to what you say, because (I don't know shit about Romanian), but I know "it hurts" is doare, from *dolet, or "want" is vrea, from *volere while forms with geminate-l may have not turned into r like this (some forms of "star" from Latin *stella have an [l], like plural stele, others seem to have lost it entirely). You should check, though anyway it's perfectly plausible.

[Thry]

Is there any evidence that allophones would become distinguished from the phoneme, and then merge with a different phoneme?

e.g. [ɾ] is an allophone of [d], then is considered distinct, then merges with [r]

Yes, Latin /s/ developed allophone [z] between vowels which then rhotacised to [ɾ] then merged with pre-existent /r/ and spread elsewhere (*flosem to florem then to flor and fills the whole paradigm).