Sound Change Quickie Thread

[jmcd]

The other solution is to have the coronals and velars palatalise in different ways, which seems fairly common.

One way is to have t>tʃ and k>c, like in the Goidelic languages.

Hypothetically, you could also have t>tɕ, k>tʃ.

[sangi39]

Another longish one from me:

Initial Phoneme Inventory

/p t k/
/m n /
/ s h/
/v r j/
/ l /

/i u/
/e o/
/ a /

Syllable structure:

CV(T)

Where C is any consonants
V is any vowel
T is one of |N| (a nasal assimilating with the following consonant for
POA) or |Q| (a “chroneme” lengthening the following consonant)

Some consonants underwent a process of lenition, where the following changes affected onsets in closed syllables:

/p/ > /b/ : /p:/ > /p/
/t/ > /d/ : /t:/ > /t/
/k/ > /g/ : /k:/ > /k/
/m/ > /m/ : /m:/ > /m/
/n/ > /n/ : /n:/ > /n/
/s/ > /z/ : /s:/ > /s/
/h/ > /?/ : /h:/ > /h/
/v/ > /v/ : /v:/ > /v/
/r/ > /r/ : /r:/ > /r/
/j/ > /j/ : /j:/ > /j/
/l/ > /l/ : /l:/ > /l/

E.G.

Code: Select all

/likæ/   > /likæ/
/likæn/  > /ligæn/
/likkæ/  > /likkæ/
/likkæn/ > /likæn/
/linkæ/  > /linkæ/
/linkæn/ > /lingæn/

Vowel length became allophonic, with short vowels appearing in closed syllables and long vowels appearing in open syllables.

E.G.

Code: Select all

/likæ/   > /likæ/   > /li:kæ:/
/likæn/  > /ligæn/  > /li:gæn/
/likkæ/  > /likkæ/  > /likkæ/
/likkæn/ > /likæn/  > /li:kæn/
/linkæ/  > /linkæ/  > /linkæ:/
/linkæn/ > /lingæn/ > /lingæn/

After this, |N| dropped out, but caused following voiceless plosives to become voiced. Geminate consonants simultaneously shortened.

E.G.

Code: Select all

/likæ/   > /likæ/   > /li:kæ:/ > /li:kæ:/
/likæn/  > /ligæn/  > /li:gæn/ > /li:gæ/
/likkæ/  > /likkæ/  > /likkæ/  > /likæ/
/likkæn/ > /likæn/  > /li:kæn/ > /li:kæ/
/linkæ/  > /linkæ/  > /linkæ:/ > /ligæ:/
/linkæn/ > /lingæn/ > /lingæn/ > /ligæ/

Vowels then underwent a number of changes:

Code: Select all

/i: y: u:/ > /ji: ju: u:/ > /ʲi: ʲu: u:/ > /ʲi ʲu u/ > /ʲi ʲu u/
/i  y  u / > / i   y  u / > / i   i  u / > / ɪ  ʊ ʊ/ > / e  o o/
/e: ø: o:/ > /je: jo: o:/ > /ʲe: ʲo: o:/ > /ʲe ʲo o/ > /ʲe ʲo o/
/e  ø  o / > / e   ø  o / > / e   e  o / > / ɛ  ɛ ɔ/ > / ɛ  ɛ ɔ/
/æ:    ɑ:/ > /ja:     a:/ > /ʲa:     a:/ > /ʲa    a/ > /ʲa    a/
/æ     ɑ / > / æ      ɑ / > / a      a / > / a    a/ > / a    a/

E.G.

Code: Select all

/likæ/   > /likæ/   > /li:kæ:/ > /li:kæ:/ > /lʲikʲa/
/likæn/  > /ligæn/  > /li:gæn/ > /li:gæ/  > /lʲiga/
/likkæ/  > /likkæ/  > /likkæ/  > /likæ/   > /leka/
/likkæn/ > /likæn/  > /li:kæn/ > /li:kæ/  > /lʲika/
/linkæ/  > /linkæ/  > /linkæ:/ > /ligæ:/  > /lʲega/
/linkæn/ > /lingæn/ > /lingæn/ > /ligæ/   > /lega/

Lastly:

Code: Select all

/e o/ > /ɪ o/
/ɛ ɔ/ > /e o/

E.G.

Code: Select all

/likæ/   > /likæ/   > /li:kæ:/ > /li:kæ:/ > /lʲikʲa/ > /lʲikʲa/
/likæn/  > /ligæn/  > /li:gæn/ > /li:gæ/  > /lʲiga/  > /lʲiga/
/likkæ/  > /likkæ/  > /likkæ/  > /likæ/   > /leka/   > /lɪka/
/likkæn/ > /likæn/  > /li:kæn/ > /li:kæ/  > /lʲika/  > /lʲika/
/linkæ/  > /linkæ/  > /linkæ:/ > /ligæ:/  > /lʲega/  > /lʲega/
/linkæn/ > /lingæn/ > /lingæn/ > /ligæ/   > /lega/   > /lɪga/

To take a more extreme example:

Code: Select all

/kykkæ/  > /kykkæ/ > /kykkæ:/ > /kykæ:/ >> /kikʲa/
/kykkæn/ > /kykæn/ > /ky:kæn/ > /ky:kæ/ >> /kʲuka/

Do these changes seem plausible?

[Nortaneous]

Glottals generally don't do that, and that last vowel change is just weird, but other than that, it looks fine to me.

[Pole, the]

The other solution is to have the coronals and velars palatalise in different ways, which seems fairly common.

One way is to have t>tʃ and k>c, like in the Goidelic languages.

Hypothetically, you could also have t>tɕ, k>tʃ.

Not just hypothetically — Polish does exactly that.

[GBR]

I feel like having uvular non-stops without other uvular consonants is pretty okay. That being said, do you mean there's no velars in your original pool except /x/? (You could also try weirder things like /xʷ/ > /ʍ/ or /w/ or /ɣ/ or /g/ or something.)

Full final phoneme inventory would be something like
/p t t^ɣ k k^ɣ/
/m n/
/r/
/ʃ ʃ^ɣ x χ/
(something something affricates something something /l/)

That'd be coming from the phonology sketch I posted here.

[sangi39]

Glottals generally don't do that

I assume we're talking about "/h/ > /?/ : /h:/ > /h/"?

and that last vowel change is just weird, but other than that, it looks fine to me.

/e o/ > /ɪ o/
/ɛ ɔ/ > /e o/

?

Yeah, I through that one in for some reason. I think it was to have a non-palatalised counterpart for /i/. Would:

/e o/ > /ɪ ʊ/
/ɛ ɔ/ > /e o/

... be more realistic, or would the change just not happen at all?

[Pazmivaniye]

Way back when I codified the sound changes from Zamétule to Proto-Maroinulic (which in turn is the parent of my primary conlang, Paleto) I had the following rules:

ps ks > ɸ x (and possibly ts > θ)
f > ɸ (this could really be anywhere, the point is just that PM had /ɸ/ but no /f/)
k > x / _y[-str] or _ɥV[-str]

Bearing in mind that there was no /u w/ and only /y ɥ/ at this stage, is it possible to justify the last change with the following?

k g > kᶣ gᶣ / _[y/ɥ]
kᶣ gᶣ > kʲ gʲ > c ɟ (this is at the same time as y ɥ > i)
c ɟ > ç g > x g

[Pole, the]

I think most probably yes, but I am drunk at the moment, so I presume you better not trust me.

[CatDoom]

Let's say you've got a language where vowel length is marginally phonemic, but long vowels only occur in open syllables, and most open syllables have long vowels. Would it be plausible for an epenthetic glottal stop to appear following short vowels, thus making vowel length an allophonic feature of vowels in open vs. closed syllables? Would the answer depend on whether or not the language in question already distinguished glottal stops as phonemic?

[Pole, the]

I think it's okay - it would be more plausible the other way 'round, but it seems okay.

But again, I am drunk, so don't trust me.

[Pazmivaniye]

Thai/Lao doesn't allow open syllables with short nuclei and adds a glottal stop.

[Qwynegold]

Can you fill in the gaps in this?

Code: Select all

p t k q -> p  t  k qʷ
? ? ?   -> pf ts x
b d ɡ   -> b  d  ɡ
? ? ?   -> bv dz ɣ

The rows don't necessarily have to correspond to each other, so for example /b d ɡ/ could come from an earlier /p t k/. I've been thinking that /q/ (or whatever the uvular protophoneme is) became labialized to distinguish it further from /k/.

The easiest thing to do would be to just have the protolang contain an aspirated and a breathy serie. But that's boring. :/ Oh, and the lack of a uvular fricative also needs to be explained.

[CatDoom]

Can you fill in the gaps in this?

Code: Select all

p t k q -> p  t  k qʷ
? ? ?   -> pf ts x
b d ɡ   -> b  d  ɡ
? ? ?   -> bv dz ɣ

The rows don't necessarily have to correspond to each other, so for example /b d ɡ/ could come from an earlier /p t k/. I've been thinking that /q/ (or whatever the uvular protophoneme is) became labialized to distinguish it further from /k/.

The easiest thing to do would be to just have the protolang contain an aspirated and a breathy serie. But that's boring. :/ Oh, and the lack of a uvular fricative also needs to be explained.

Did the protolang allow onset clusters? Maybe it had a uvular fricative that merged with x, and subsequently x assimilated to the PoA and voice of a preceding stop in clusters. I know that x > f is an attested sound change (there seems to be an odd connection between velar and labial consonants), and I'm fairly sure x > s next to coronals is an attested example of assimilation. /kx/ and /gɣ/ are fairly unusual sounds, so it doesn't seem unreasonable for them to lenite to /x/ and /ɣ/, possibly as a result of areal influences from languages without velar affricates. The changes would be something like this:

x̠ > x
px > pf
bx > bv
tx > ts
dx > dz
kx > x
gx > gɣ > ɣ

[Pole, the]

Can you fill in the gaps in this?

Code: Select all

p t k q -> p  t  k qʷ
? ? ?   -> pf ts x
b d ɡ   -> b  d  ɡ
? ? ?   -> bv dz ɣ

The rows don't necessarily have to correspond to each other, so for example /b d ɡ/ could come from an earlier /p t k/. I've been thinking that /q/ (or whatever the uvular protophoneme is) became labialized to distinguish it further from /k/.

The easiest thing to do would be to just have the protolang contain an aspirated and a breathy serie. But that's boring. :/ Oh, and the lack of a uvular fricative also needs to be explained.

1. /pf ts x bv dz ɣ/, earlier /pf ts kx bv dz ɡɣ/, could come from geminate /pː tː kː bː dː ɡː/ here.

2. /pf ts x bv dz ɣ/ could come from palatalized /p t k b d ɡ/ — it would also explain, why /q/ hasn't got its counterpart. One of the routes could be e.g. /pʲ tʲ kʲ bʲ dʲ ɡʲ/ → /pç tç cç bʝ dʝ ɟʝ/ → /pʃ tʃ ʃ bʒ dʒ ʒ/ → /px ts x bɣ dz ɣ/.

3. T'was other way 'round. The initial contrast was /p t k : b d ɡ/, but then it shifted via /pʰ tʰ kʰ p t k/ into /pf ts x p t k/. The voicing distinction is a feature that appeared independently.

3a. One example would be oral /p t k b d ɡ/ and prenasalized /mp nt ŋk mb nd ŋɡ/ first shifting via /pʰ tʰ kʰ p t k/ and /mpʰ ntʰ ŋkʰ mp nt ŋk/ into /pf ts x p t k mpf nts ŋx mp nt ŋk/ and then going the Greek/Hungarian style: /mpf nts ŋx mp nt ŋk/ → /bv dz ɣ b d ɡ/.

[Nortaneous]

light/dark contrast realized alternately as palatalization or absence of velarization

p pʷ tʲ t k kʷ > p pf ts t k qʷ
b bʷ dʲ d g gʷ > b bv dz d g ɢʷ
ɢʷ > ʁʷ > ʁ > ɣ

or prevelar/postvelar instead of k/kw

anyway how'd those bantu langs do it

[Rungari]

How realistic is the "hardening" of /β ð ɣ̞/ back into /b d g/? I think I read this happened in Germanic somewhere, but I can't remember where.

[Nortaneous]

θ d and ɣ g are both attested in Germanic.

[WeepingElf]

How about the following?

as > i /_#
ax > u /_#

Phoneme inventory before the change:

/pʰ tʰ kʰ kʷʰ/
/p t k kʷ/
/b d g gʷ/
/s x xʷ/
/m n/
/l r/
/w j/
/i a u/

[Cedh]

How about the following?

as > i /_#
ax > u /_#

Should be possible with intermediate changes roughly like this:

V > Və̯ / _[+fricative]#
aə̯ > aɯ̯ / _[+velar]
aə̯ > aɪ̯ / elsewhere
[+fricative] > Ø / V_#
aɪ̯ aɯ̯ ( > ɛː ɔː > eː oː > iː uː ) > i u

[WeepingElf]

How about the following?

as > i /_#
ax > u /_#

Should be possible with intermediate changes roughly like this:

V > Və̯ / _[+fricative]#
aə̯ > aɯ̯ / _[+velar]
aə̯ > aɪ̯ / elsewhere
[+fricative] > Ø / V_#
aɪ̯ aɯ̯ ( > ɛː ɔː > eː oː > iː uː ) > i u

Yes, something like that I have been thinking about.

[Richard W]

How realistic is the "hardening" of /β ð ɣ̞/ back into /b d g/? I think I read this happened in Germanic somewhere, but I can't remember where.

It looks as though it happened in Latin except word-initially.

The hardening happened word-initially and after nasals in Proto-Germanic > English, though /ɣ/ hardened latest word initially - possibly not before the Norman conquest. The hardening was more thorough in German (English 'wife' v. German 'Weib'), though /ɣ/ didn’t harden to a stop in Dutch, so I don't know where the hardening/non-hardening boundary lay.

[Hallow XIII]

This is one of those sound changes where idiosyncrasy may be expected. Some dialects of Scottish Gaelic harden final /G/ when it occurs as part of the Verbal Noun marker -adh (e.g. a' chreachadh [ə ˈxɾʲɛxək]), others leave it unchanged or delete it (Irish, for comparison, generally shifted -adh to -ú, which is also spelt this way when the affix is used derivationally). I think there may be dialects that consistently harden it in all final positions but I couldn't put my finger on any at the moment.

[Qwynegold]

Did the protolang allow onset clusters? Maybe it had a uvular fricative that merged with x, and subsequently x assimilated to the PoA and voice of a preceding stop in clusters. I know that x > f is an attested sound change (there seems to be an odd connection between velar and labial consonants), and I'm fairly sure x > s next to coronals is an attested example of assimilation.

Yes, and the daughterlang still has some quite complex clusters, so I don't know how I would explain only clusters with /x/ simplifying when there are other, more complicated things left.

1. /pf ts x bv dz ɣ/, earlier /pf ts kx bv dz ɡɣ/, could come from geminate /pː tː kː bː dː ɡː/ here.

2. /pf ts x bv dz ɣ/ could come from palatalized /p t k b d ɡ/ — it would also explain, why /q/ hasn't got its counterpart. One of the routes could be e.g. /pʲ tʲ kʲ bʲ dʲ ɡʲ/ → /pç tç cç bʝ dʝ ɟʝ/ → /pʃ tʃ ʃ bʒ dʒ ʒ/ → /px ts x bɣ dz ɣ/.

3. T'was other way 'round. The initial contrast was /p t k : b d ɡ/, but then it shifted via /pʰ tʰ kʰ p t k/ into /pf ts x p t k/. The voicing distinction is a feature that appeared independently.

3a. One example would be oral /p t k b d ɡ/ and prenasalized /mp nt ŋk mb nd ŋɡ/ first shifting via /pʰ tʰ kʰ p t k/ and /mpʰ ntʰ ŋkʰ mp nt ŋk/ into /pf ts x p t k mpf nts ŋx mp nt ŋk/ and then going the Greek/Hungarian style: /mpf nts ŋx mp nt ŋk/ → /bv dz ɣ b d ɡ/.

Gemination word-initially? I don't know...

light/dark contrast realized alternately as palatalization or absence of velarization

p pʷ tʲ t k kʷ > p pf ts t k qʷ
b bʷ dʲ d g gʷ > b bv dz d g ɢʷ
ɢʷ > ʁʷ > ʁ > ɣ

or prevelar/postvelar instead of k/kw

anyway how'd those bantu langs do it

Hmm, this and Pole's second suggestion made think about having a pharyngealization contrast. Maybe C[-voiced]ˤ could turn into Cx and C[+voiced]ˤ into Cɣ? Oh, but then I run into the same problem that CatDoom's suggestion had. :/

Okay, yet another idea I came up with thanks to Pole's third suggestion. What if I had a three-way contrast of voiceless, slack voice and stiff voice for plosives and fricatives? Would that, and these changes be plausible?:
t d̥ d̬ tʰ d ɗ t͡s t d
s z̥ z̬ s z ɗ͡z s z d͡z

[Click]

Are the following sound changes plausible?

1. [pt] [kt] → [tʂ] [tɕ]
2. {[w] [j]} → [h] / C_
3. Vj Vw → jV wV

[Nortaneous]

1. sure, maybe with intermediate step of tɕʷ
2. no
3. no, though chechen has diphthong metathesis after pharyngealized consonants

word-initial geminates are fine, see http://www.linguistics.ucla.edu/afla12/ ... B_1200.pdf