Sound Change Quickie Thread

[Hallow XIII]

Yes, this happens a lot. Also assibilation is basically the favorite sound change for anything palatal so I advise you not to listen to sirdanilot.

[communistplot]

What would you guys do with this mess?

Phonotactics
Syllable Structure
Syllable structure is maximally (C)(C)(G)V(C) tho typical syllables tend toward CV or CGV with C being any consonant, G any glide (or /ɬ ɬʰ/) or liquid and V any vowel.

Root Structure
Roots are predominately monosyllabic and biliteral tho there are bi- and tri-syllabic roots with tri- and quadriliteral structures found as well.

Allowed Clusters
Initial
• Any consonant
• Any consonant minus Liquids or Glides (or /tɬ tɬʰ/) followed by a Liquid or Glide, or /ɬ ɬʰ/
• Any of the above minus Fricatives if preceded by /s sʰ/
• Any nasal followed by its homorganic plosive, either fortis or lenis
• /n/ followed by /s/ or /n s sʰ/ followed by /t͡ɬ t͡ɬʰ/
Medial
• Any cluster allowed initially
• Any liquid or glide (or /ɬ ɬʰ/) followed by a consonant and either a vowel or another liquid or glide (or /ɬ ɬʰ/)
Final
• Any single consonant
Hiatus
Hiatus is not allowed and vowels which would occur in such a situation do one of two things, they assimilate and form a diphthong or the second element drops, this occurs even over morpheme boundaries and with regards to the second over lexical boundaries as well.

Allophony
• Lenis plosives become unreleased syllable finally and fortis plosives become lenis.
• Progressive assimilation occurs when a lenis or fortis consonant begins a cluster assimilating to either depending on whichever begins the cluster (e.g. /skʰā/ becomes [skā] and /sʰkā/ becomes [sʰkʰā]).
• /k kʰ ŋ/ become [c cʰ ɲ] when preceding any of /ɹʲ j i ɪ/.
• Likewise /ɹʲ/ drops before /j/ but becomes /j/ when before /i ɪ/.
• /ɹʲ/ vocalises to [ɚ̯] postvocalically at the end of a word or at the end of a syllable before another consonant when following a high vowel, otherwise it is dropped.
• Likewise /l/ becomes [w̯] when following any vowel (except intervocalically) other than /u o/ which it merges with.
• /l/ becomes [ʟ] when following /k kʰ ŋ/.
• /ɰ j/ become [ɡ ɟ] intervocallically, occurring across morpheme and lexeme boundaries.
• If two syllables with the same tone are next to each other the second one will shift (rising to falling, middle to checked, checked to falling, falling to middle).
• A vowel followed by /ʔ/ has automatic checked tone (e.g. /ǎ ʔê/ becomes [aʔ ʔê]).
• Checked tone becomes middle tone before nasals.

[StrangerCoug]

I have a few issues with the transcription:

• I'm not a big fan of <ii> for /i/ or <uu> for /u/, but the only alternative coming to mind essentially uses English spelling rules instead (<ee> for /i/ and <oo> for /u/), and I'm not sure that makes the transcription all that interesting. MAYBE you can justify your system by saying a long-short distinction got converted into a tense-lax distinction, but you still have the issue of the closing diphthongs looking ugly.
• <ñ> is more typically used for /ɲ/, not /ŋ/ (though a Wikipedia search says it's used for the latter in Crimean Tatar).
• You use <ḳ> but not <k>—why?

[Nortaneous]

<ñ> /ŋ/ is also used in Turkmen (and maybe some other Turkic languages?) and some Latin transcriptions of Chukchi.

[communistplot]

I have a few issues with the transcription:

• I'm not a big fan of <ii> for /i/ or <uu> for /u/, but the only alternative coming to mind essentially uses English spelling rules instead (<ee> for /i/ and <oo> for /u/), and I'm not sure that makes the transcription all that interesting. MAYBE you can justify your system by saying a long-short distinction got converted into a tense-lax distinction, but you still have the issue of the closing diphthongs looking ugly.
• <ñ> is more typically used for /ɲ/, not /ŋ/ (though a Wikipedia search says it's used for the latter in Crimean Tatar).
• You use <ḳ> but not <k>—why?

I like my transcription. I think my transcription gives it a certain feel for me that a different transcription scheme wouldn't, y'know?

But <ḳ> is an error from when I was transcribing aspirates with dotted consonants, it should be <k>.

Anyway, I want to see what interesting sound changes can be done.

[Hallow XIII]

if tK tK_h pattern as clusters in your language why do you analyze them as separate phonemes

[communistplot]

if tK tK_h pattern as clusters in your language why do you analyze them as separate phonemes

More an oversight when I was writing the cluster rules. /t͡ɬ/, /t͡ɬʰ/ is phonemic, not a cluster.

[Qwynegold]

You know, even /aʔ/ needs to have a tone phonetically, unless it's voiceless. If you're looking for SC suggestions, here's a couple:
ɹʲ ʝ ʒ z or something like that
ɰ w (merging with /l/ in some cases)
ea eː
oa oː

[StrangerCoug]

What can I do with /tʃʰ/ that's consistent with /pʰ tʰ kʰ/ → /ɸ θ x/?

[Matrix]

/ʃ/?

[StrangerCoug]

Possible idea.

Another question: Is /ŋk/ → /ŋɡ/ → /ɲɟ/ → /ndʒ/ plausible in the environment F_V (F being a front vowel, V being any vowel at all)?

[8Deer]

Possible idea.

Another question: Is /ŋk/ → /ŋɡ/ → /ɲɟ/ → /ndʒ/ plausible in the environment F_V (F being a front vowel, V being any vowel at all)?

Seems plausible. I'd expect /mp/ and /nt/ to shift to /mb nd/ as well.

[communistplot]

You know, even /aʔ/ needs to have a tone phonetically, unless it's voiceless. If you're looking for SC suggestions, here's a couple:
ɹʲ ʝ ʒ z or something like that
ɰ w (merging with /l/ in some cases)
ea eː
oa oː

/aʔ/ is tonemically middle, with a glottal release. And this was more an experiment, just wondering what others would do to that phonology. I do like the idea of /ɹʲ/ becoming /z/ tho, in my notes I have it merging with /j/ ala Burmese.

[GreenBowTie]

What would you guys do with this mess?

i'd probably start with screwing with the vowels. tbh i don't know anything about tones, so maybe nothing i'm proposing is even possible with them, but this is along the lines of how i might start:

step one: closing diphthongs become a semivowel plus a vowel
ia, ea ja
iu ju
oa, ua wa
ui wi

step two: remaining diphthongs become monophthongs
ai ɛ
au ɔ
ei e
ou o

this leaves us with just monophthongs, and also produces the new vowel phonemes /ɛ/ and /ɔ/ (which complement i/ɪ and u/ʊ) as well as the new consonant phoneme /w/ (which could end up merging with /ɰ/).

next you could try umlaut or a similar process. this could be triggered by when a back vowel is followed by a front vowel, or maybe by one of the palatals. unstressed vowels (including those triggering umlaut) could then be reduced or even deleted, giving new phonemes

the obvious route for the consonants is to simply turn it into a voiced distinction: p pʰ p b. it might be more fun, though, to turn aspirated consonants (at least the stops) into affricates, which might then devolve into fricatives. you could then introduce a voicing contrast by voicing consonants intervocalically and then in certain environments deleting unstressed vowels. having your only rhotic be palatal seems pretty likely to change as well: it could turn into a simple, non-palatal rhotic, or maybe merge with /j/ or something. i'd definitely save that change for after you've applied umlaut though.

i'll try to come up with more when the computer lab's not about to close. hope at least some of this is helpful

[Tiamat]

would this be reasonable?

wə(:) > o(:)
jə(:) > e(:)

while following stay the same?
wa(:)
ja(:)
a(:)w$
a(:)j$

There is no native ə(:)w$ or ə(:)j$, btw.

I was thinking about having some intermediate steps of
wə(:) > wo(:) > (ban on /w/ before back rounded vowels) o(:)
jə(:) > je(:) > (ban on /j/ before front vowels) e(:)

[----]

You don't even need intermediate steps. That's perfectly fine. But if you're not sure you could always have regular /@/ elided in other situations, or shifted to some other vowel or something.

[KathTheDragon]

Yeah, that's totally fine.

[Nortaneous]

What would you guys do with this mess?

c cʰ > ʃ ʃʰ
ʟ > w
w ɰ j > b ɟ g / V_V
fricative aspiration distinction lost syllable-finally
ns > nz > nd > d
vowel loss that makes b ɟ g phonemic
ɰ merges with the ɚ allophone of the rhotic as ʕ

[Dē Graut Bʉr]

w ɰ j > b ɟ g / V_V

Don't you mean j w > ɟ g?

[Qwynegold]

w ɰ j > b ɟ g / V_V

Don't you mean j w > ɟ g?

He probably meant w ɰ j > b g ɟ. :S

[Nortaneous]

right, have the existing glide fortition process expand to /w/

[----]

This seems like a nice time to ask about a similar set of changes in my Austronesian conlang. I have intervocalic glides fortitiating like this:

w j > b dʒ

but I was wondering if I could also have a change that happens about the same time where word initial bə bu > o u as a sort of hypercorrection type thing I guess?

for example:
*beRas > *bəgah > ogɪh
*buaq > *buwaq > *bəwa > *bəba > oba
*buliR > *bulij > uləj

[CatDoom]

That's about the opposite of what one would normally expect to happen; in general, word-initial consonants are in a stronger position than intervocalic consonants. You may be able to find an exception, but I tend to think that if glides undergo fortition between vowels, stops are unlikely to weaken to zero initially.

[Nortaneous]

Changes like that happen sometimes in Austronesian languages.

KQ archive says Polynesian has w > p. PMP -> Proto-Oceanic neutralized voice contrast and redeveloped it from prenasalization -- p b > p, mp mb > b -- so could've gone through an intermediate stage, w > b > p. Levei and Drehet had w j > p. You get a few other strange fortition-like sound changes too, so for an Austronesian conlang, intervocalic glide fortition isn't implausible.

Did that change happen anywhere? I thought it did, but can't find it now.

As for word-initial: probably wouldn't be hypercorrection since different environment. Say b > bʷ before ə u. (Assuming you don't have /o/ at that stage.) Probably its sound alike, since fortition happens -- might want to add some general confusion there, say a few words jump between the two irregularly -- so bʷ > w as w > b, #w merges with following vowel, whatever. Or bʷ could just transfer [+labial] to the following nonfront vowel and drop at a totally different time.

(If you want to add front rounded vowels, that could be a good way to do it -- then once you've got ö y from #ba #bi or whatever, vowel harmony or Souletin(?) Basque-like assimilation effects, hypothetical *bilin > ülün, whatever.)

[----]

Did that change happen anywhere? I thought it did, but can't find it now.

w j > b dʒ and variants of it happened in several languages of Borneo.

I'm actually kind of interested in that b > bʷ thing. Maybe I could have all the labials do it and have other changes, so like
bʷ pʷ mʷ > w f m
or maybe mʷ > ŋ instead. I think that happened in the Oceanic branch.

I dunno. I just don't want half the language to be bəbəb bə bəbəbbb because sequences like *buV are quite common in PAN.