Zompist Bboard Again

Knit Tie wrote: ↑Sat Dec 28, 2019 11:31 amAlso, can
ʃp ʃt ʃk → f ʃt x, i.e. can the middle cluster remain as the other two simplify?

This is basically attested from Albanian: word-initial *sp *st *sk > /f ʃt h/. Some sources suggest that there was an intermediate step with metathesis to *ps *ks, another route could be debuccalization or aspiration spreading through *spʰ *stʰ *skʰ. (Compare Middle Indic: /sp st sc sk/ > initial /pʰ tʰ cʰ kʰ/, medial /ppʰ ttʰ ccʰ kkʰ/.)

Also, digging up some older questions:

Knit Tie wrote: ↑Thu Nov 28, 2019 3:18 pmYod becoming a postalveolar fricative during devoicing isn't unrealistic, is it? I just can't find any examples of it at all.

Closest I know of is Mator, where *Vjt > Vst, entirely possibly through *Vʃt (just this one cluster though; no change in *Vjʔ at least).

bradrn wrote: ↑Thu Nov 14, 2019 5:29 amWikipedia lists Kodagu (a.k.a Kodava) as having this contrast [all of /i ɨ ɯ u/] as well

Kodagu has both [ɨ ɯ], but they don't contrast. [ɨ] is an epenthetic unstressed vowel and only /ɯ/ is phonemic.

As far as phonological representation goes, in principle /ɨ ɯ/ can be still both represented with just [±front ±labial], via underspecification: one of them as [-front] (or [+back]) with no labiality specified, the other as [-labial] with no frontness specified.

A third simple possibility is that in several languages, what gets called "/ɨ/" is phonologically [-high] (or at least not specified as [+high]) and could be also described as /ə/ with a relatively high realization.

I have these nonconcatenative patterns for various triliteral nouns:
https://conworkshop.com/view_article.ph ... e1b424a05e

If one were to reconstruct what morphology/morphemes lead to these patterns, what do you think they could be?

Can /ç/ become /s/ instead of /ʃ/ or /ç/?

Knit Tie wrote: ↑Wed Jan 08, 2020 5:54 pm Can /ç/ become /s/ instead of /ʃ/ or /ç/?

Note that ç > ʃ > s is definitely possible.

Knit Tie wrote: ↑Wed Jan 08, 2020 5:54 pm Can /ç/ become /s/ instead of /ʃ/ or /ç/?

depends on what /ʃ/ is, I think - ç > ɕ > s̻, sure, if ʃ is apical

Nortaneous wrote: ↑Wed Jan 08, 2020 7:01 pm

Knit Tie wrote: ↑Wed Jan 08, 2020 5:54 pm Can /ç/ become /s/ instead of /ʃ/ or /ç/?

depends on what /ʃ/ is, I think - ç > ɕ > s̻, sure, if ʃ is apical

ʃ is absolutely apical postalveolar in my lang, yes.

Is /ŋ/ > /m/ (unconditionally) possible? It can happen with rounded ŋʷ, but I can't find any examples of plain /ŋ/. If that's not possible, what are some other ways in which /ŋ/ could evolve?

The consonant inventory is /n ŋ t k s r/

Darren wrote: ↑Thu Jan 09, 2020 4:07 pm Is /ŋ/ > /m/ (unconditionally) possible? It can happen with rounded ŋʷ, but I can't find any examples of plain /ŋ/. If that's not possible, what are some other ways in which /ŋ/ could evolve?

The consonant inventory is /n ŋ t k s r/

That sounds suspicious. I would say /ŋ/ → /g/ or /ŋ/ → /n/ is much more plausible.

(By the way, that’s a pretty interesting consonant inventory you’ve got there. How come you have so few consonants?)

bradrn wrote: ↑Thu Jan 09, 2020 4:11 pm

Darren wrote: ↑Thu Jan 09, 2020 4:07 pm Is /ŋ/ > /m/ (unconditionally) possible? It can happen with rounded ŋʷ, but I can't find any examples of plain /ŋ/. If that's not possible, what are some other ways in which /ŋ/ could evolve?

The consonant inventory is /n ŋ t k s r/

That sounds suspicious. I would say /ŋ/ → /g/ or /ŋ/ → /n/ is much more plausible.

I know it sounds suspicious. I'm planning to make a few descendants, which treat /ŋ/ differently so they look more distinctive. I might be able to have something like /ŋu/ > /ŋʷu/ > /mu/ while elsewhere there's another change... maybe to /j/ before unrounded vowels.

(By the way, that’s a pretty interesting consonant inventory you’ve got there. How come you have so few consonants?)

Because I like it! It's not very naturalistic but I prefer small consonant inventories. Just a matter of aesthetics.

Darren wrote: ↑Thu Jan 09, 2020 4:07 pm Is /ŋ/ > /m/ (unconditionally) possible? It can happen with rounded ŋʷ, but I can't find any examples of plain /ŋ/. If that's not possible, what are some other ways in which /ŋ/ could evolve?

The consonant inventory is /n ŋ t k s r/

Given such a tiny inventory, I'd say it looks reasonable. The stops can be categorised as [+/-]nasal and [+/-]coronal. Some Romanian examples spring to mind: Latin lignum > Rom. lemn 'wood' and, non-nasally Latin nebula > Rom negură 'fog'. The first change is a parallel to Latin 'ct' > Romanian 'pt', so not completely unconditioned, and the second change is sporadic.

A similar shift is Tai Lue /nɯŋ/ > /m̩/ 'one'; this is not a placeless nasal.

Richard W wrote: ↑Thu Jan 09, 2020 6:05 pm

Darren wrote: ↑Thu Jan 09, 2020 4:07 pm Is /ŋ/ > /m/ (unconditionally) possible? It can happen with rounded ŋʷ, but I can't find any examples of plain /ŋ/. If that's not possible, what are some other ways in which /ŋ/ could evolve?

The consonant inventory is /n ŋ t k s r/

Given such a tiny inventory, I'd say it looks reasonable. The stops can be categorised as [+/-]nasal and [+/-]coronal. Some Romanian examples spring to mind: Latin lignum > Rom. lemn 'wood' and, non-nasally Latin nebula > Rom negură 'fog'. The first change is a parallel to Latin 'ct' > Romanian 'pt', so not completely unconditioned, and the second change is sporadic.

A similar shift is Tai Lue /nɯŋ/ > /m̩/ 'one'; this is not a placeless nasal.

That makes sense... speakers would probably want to distinguish /ŋ/ from /k/ (likely [k~g]) more, and under a [±coronal] analysis they could switch to labials. Would this mean that having /k/ as [p~k] dialectally or allophonically would be reasonable?

That would certainly be possible, since both are [+grave] (and there's precedent for the mirror-image unconditioned sound change *p --> k in Arapahoan, and intervocalic *p --> kʷ in Wichita) and of course in very small consonant inventories like this there's often some pretty large allophonic or free variation in realization of the phonemes

Whimemsz wrote: ↑Fri Jan 10, 2020 12:31 pm That would certainly be possible, since both are [+grave] (and there's precedent for the mirror-image unconditioned sound change *p --> k in Arapahoan, and intervocalic *p --> kʷ in Wichita) and of course in very small consonant inventories like this there's often some pretty large allophonic or free variation in realization of the phonemes

Thanks! That's helpful, but I'm not sure if I understand the term "[±grave]". Is there a specific linguistic meaning for this?

Yes, [+grave] consonants are labials and velars, which have some acoustic/perceptual similarities and which not infrequently pattern as a single class or easily interchange in phonological rules or sound changes. (A few examples of such sound changes -- Arapahoan *p --> k and Romanian velar --> labial before some consonants -- have already been given, and there's also, e.g., English sporadic x# --> f# as in "laugh".)

often called 'peripheral' in Australian languages where the category is relevant, which is apparently common for them

Whimemsz wrote: ↑Fri Jan 10, 2020 7:33 pmthere's also, e.g., English sporadic x# --> f# as in "laugh".)

That only occurs after historical /w/: lawx > laxʷ > laf

Thanks for the help everyone. I'll probably end up having *ŋ reflected as /m/ and *k as /k/, realised as [p] before rounded vowels or consonants. I'd be suspicious of a sound change that completely removed the only velar for no strong reason.

KathTheDragon wrote: ↑Fri Jan 10, 2020 11:09 pm

Whimemsz wrote: ↑Fri Jan 10, 2020 7:33 pmthere's also, e.g., English sporadic x# --> f# as in "laugh".)

That only occurs after historical /w/: lawx > laxʷ > laf

Oh. Hm, yes. Good point! (Anyway the statement about grave consonants often patterning together or turning into one another is still true.)

Does anyone know of substantiations/justifications of /q/ > /h/ occuring? Either direct or via an intermediary, or restricted to certain environs (I'm thinking word-finally is probably a sure bet).

holbuzvala wrote: ↑Mon Jan 13, 2020 5:29 am Does anyone know of substantiations/justifications of /q/ > /h/ occuring? Either direct or via an intermediary, or restricted to certain environs (I'm thinking word-finally is probably a sure bet).

Proto-Malayo-Polynesian to Proto-Chamic and Proto-Malayo-Javanic. q → χ is attested in Abzhywa and some other Abazgi languages, Ayachuco and Palauan, from there it's easy to go to /h/. That's all to be taken with a grain of salt, though, as often back-consonants are hard to reconsrtuct.