Zompist Bboard Again

According to wikipedia, some Somali dialects reflect standard /q/ as /x/. It looks like this claim comes from Saeed (1993), which I don't have access to, but the article cites another paper, Gabbard (2010), which claims that q > ʁ / V_V allophonically. From there you could easily get to [h]. Also, according to Fre Woldu (1985), in Tigrinya there's a synchronic rule k k' kʷ kʷ' > X X' Xʷ Xʷ' /V_ (though IIRC some other sources claim > x x' xʷ xʷ'), so it could certainly happen with [q], which is much less stable than [k].

Georgian reflects Proto-Kartvelian *q(ʰ) as /x/, while maintaining the ejective /q'/.

I don't know whether we reconstruct aspiration back to PK, though--what are the phonetics in Laz/Svan/Mingrelian?

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).

PMP *q > Proto-Malayan *h > Malay h

Allophonic [ꭓ] or [ʁ] for /q/ is common; from Brazilian Portuguese we know these can shift to [h]. For example, Tehit:
/saq/ [sɑqꭓ ~ sɑq]
/tqoq/ [təɢɔqꭓ ~ təgɔq]
/msqeq/ [msqɛqꭓ ~ msqɛq]

/qakm/ [qꭓɑkm ~ ꭓɑkm]
/qej/ [qꭓej ~ ꭓej]

/taqa/ [tɑɢɑ ~ tɑʁɑ]
/maaqo/ [maːɢɔ ~ maːʁɔ]
/wqaq/ [wəɢɑq ~ wəʁɑq]
/mqieq/ [məɢɪjɛqχ ~ məʁɪjɛqχ]

/mqlaa/ [mqlaː ~ mɢlaː]
/wqro/ [wqrɔ ~ wɢrɔ]
/fqndan/ [ɸqndan ~ ɸɢndan]

/tsqaat/ [tsꭓaːtʰ]
/mqan/ [mχɑn]
/fqmŋga/ [fꭓəmŋgɑ] (but /mqmŋga/ [məqmŋgɑ])

Is post-/ɑ/ uvularization of velars a thing? I find a word like Vrkhazhian /ɑmɑːk/ "I, me" naturally becomes [ɑmɑːq]

Ahzoh wrote: ↑Mon Jan 13, 2020 12:47 pm Is post-/ɑ/ uvularization of velars a thing? I find a word like Vrkhazhian /ɑmɑːk/ "I, me" naturally becomes [ɑmɑːq]

I do not have the time at the moment to look up documentation of this change per se, but it seems perfectly plausible to me.

What sound changes or allophony could I do to establish a natural class distinction between peripheral consonants (bilabial and velar) and central consonants (central alveolar and lateral alveolar)?

I thought an assimilation of stops or fricatives or both, but the phonology has gaps such that there are no bilabial counterparts to the velar fricatives and no lateral counterparts to the central alveolar stops. Also /pk bg kp gb/[k: g: p: b:] is kinda ugly and feels like it would lead to more homonyms than I'd like.

Ahzoh wrote: ↑Mon Jan 13, 2020 1:57 pm What sound changes or allophony could I do to establish a natural class distinction between peripheral consonants (bilabial and velar) and central consonants (central alveolar and lateral alveolar)?

Distribution. In some Australian languages, peripherals are strongly preferred to coronals in word-initial position, and in some (all?) Germanic languages, coronals are strongly preferred to peripherals in unstressed syllable codas.

Allophony. I can't think of precedent off the top of my head, but /ə/ could be realized as [ʊ] before peripherals, or peripheral plosives could lenite to fricatives intervocalically, or something.

Recent sound changes. In Western Nga'o, mb nd ŋg > p nd k.

Nortaneous wrote: ↑Mon Jan 13, 2020 2:10 pmDistribution. In some Australian languages, peripherals are strongly preferred to coronals in word-initial position, and in some (all?) Germanic languages, coronals are strongly preferred to peripherals in unstressed syllable codas.

Yes, it seems subconsciously I already have it that velars and bilabials are most often in word-initial or medial position and least often in word-final position. I dunno, I just hate word-final /p b/.

Allophony. I can't think of precedent off the top of my head, but /ə/ could be realized as [ʊ] before peripherals, or peripheral plosives could lenite to fricatives intervocalically, or something.

I don't have a schwa but maybe the consonant class could determine whether the epenthetic vowel is /e/ or /A/.

Could the liquids /j w r/ do any special interactions based on these classes?

Ahzoh wrote: ↑Mon Jan 13, 2020 12:47 pm Is post-/ɑ/ uvularization of velars a thing? I find a word like Vrkhazhian /ɑmɑːk/ "I, me" naturally becomes [ɑmɑːq]

i think German also shifts its /x/ to uvular when following /A/. (But not t he stops)

Ahzoh wrote: ↑Mon Jan 13, 2020 12:47 pm Is post-/ɑ/ uvularization of velars a thing? I find a word like Vrkhazhian /ɑmɑːk/ "I, me" naturally becomes [ɑmɑːq]

Hewa: "The velar consonants are backed further when preceded or followed by a low or back vowel."

Duwet: " /k/: /k/ is a back velar [ḳ] after /a,ia,ea/."

As I've mentioned here before, a lot of Japanese speakers seem to have k > q /_a at least in certain informal contexts

edit: also apparently Multicultural London English

Darren wrote: ↑Thu Jan 09, 2020 4:07 pmwhat are some other ways in which /ŋ/ could evolve?

A few attested paths from Uralic for getting rid of *ŋ:

(1) Vocalization
Finnic: *ŋ > *w next to rounded vowels, *ŋ > ∅ otherwise
Mordvinic: *ŋ > *w (> v) next to back vowels, *ŋ > j next to front vowels

(2) POA changes
Komi: *ŋ > m next to rounded vowels, *ŋ > ɲ next to front vowels, *ŋ > n otherwise
dialectal Udmurt: *ŋ > ɲ next to front vowels, *ŋ > n otherwise (so same as Komi but no > m)
dialectal Northern Sami: *ŋ > ɲ always (merger with pre-existing /ɲ/)

General *ŋ > m sounds possible as a merger, but seems weird if the language previously lacks /m/.

Most Mayan languages also had unconditional *ŋ --> x, unconditional *ŋ --> h, or unconditional *ŋ --> n.

Maori dialects: ŋ > k (South Island), ŋ > n (Tuhoe).

Looking at ABVD, some Ivatan languages have forms like tadinya for *Caliŋa, and Tolomako has tsaliɣa-ku. The Ivatan form might be palatalization, but ŋ > ɣ seems reasonable.

Tropylium wrote: ↑Tue Jan 14, 2020 1:00 pm

Darren wrote: ↑Thu Jan 09, 2020 4:07 pmwhat are some other ways in which /ŋ/ could evolve?

A few attested paths from Uralic for getting rid of *ŋ:

(1) Vocalization
Finnic: *ŋ > *w next to rounded vowels, *ŋ > ∅ otherwise
Mordvinic: *ŋ > *w (> v) next to back vowels, *ŋ > j next to front vowels

(2) POA changes
Komi: *ŋ > m next to rounded vowels, *ŋ > ɲ next to front vowels, *ŋ > n otherwise
dialectal Udmurt: *ŋ > ɲ next to front vowels, *ŋ > n otherwise (so same as Komi but no > m)
dialectal Northern Sami: *ŋ > ɲ always (merger with pre-existing /ɲ/)

General *ŋ > m sounds possible as a merger, but seems weird if the language previously lacks /m/.

Thanks! The Komi change is really similar to what I had in mind. I know that ŋ > m unconditionally would be wierd, I might add that in just one dialect as a result of influence from an unrelated language. The other changes are interesting too, there are a couple of languages planned for this branch.

Nortaneous wrote: ↑Tue Jan 14, 2020 2:33 pm Maori dialects: ŋ > k (South Island), ŋ > n (Tuhoe).

Looking at ABVD, some Ivatan languages have forms like tadinya for *Caliŋa, and Tolomako has tsaliɣa-ku. The Ivatan form might be palatalization, but ŋ > ɣ seems reasonable.

I'd thought of doing ŋ > k (most likely ŋ > g > k), but I didn't know there was a natlang precedent. Maybe [ɣ] could surface as an allophone.

Whimemsz wrote: ↑Tue Jan 14, 2020 1:19 pm Most Mayan languages also had unconditional *ŋ --> x, unconditional *ŋ --> h, or unconditional *ŋ --> n.

I saw those in the Diachronica and I was a bit surprised by ŋ → x.

Thanks for all the ideas, I think I'll have to have quite a few languages in this branch to explore all the possibilities.

Nortaneous wrote: ↑Mon Jan 13, 2020 9:07 amAllophonic [ꭓ] or [ʁ] for /q/ is common

I'm guessing it's just a mistake, but you used U+AB53 for chi there. IPA [X] is supposed to simply be U+3C7 GREEK SMALL LETTER CHI, very commonly covered in fonts. U+AB53 is a Latin Extended-E symbol specific to German dialectology to represent palatal [ç], and it is very rarely covered in fonts. (I can't actually see your small chi on default latest Ubuntu.)

Ser wrote: ↑Tue Jan 14, 2020 5:28 pm

Nortaneous wrote: ↑Mon Jan 13, 2020 9:07 amAllophonic [ꭓ] or [ʁ] for /q/ is common

I'm guessing it's just a mistake, but you used U+AB53 for chi there. IPA [X] is supposed to simply be U+3C7 GREEK SMALL LETTER CHI, very commonly covered in fonts. U+AB53 is a Latin Extended-E symbol specific to German dialectology to represent palatal [ç], and it is very rarely covered in fonts. (I can't actually see your small chi on default latest Ubuntu.)

Well, technically speaking, I believe you’re supposed to use U+AB53 for IPA. Like using Latin gamma ⟨ɣ⟩ instead of Greek gamma ⟨γ⟩. But as you note, few fonts support it, and most IPA keyboards don’t use it, so usually the Greek letter is used.

(It would be nice if more fonts supported it though: sometimes chi is homoglyphic with ⟨x⟩, and using Latin chi avoids this problem…)

bradrn wrote: ↑Tue Jan 14, 2020 5:35 pmWell, technically speaking, I believe you’re supposed to use U+AB53 for IPA. Like using Latin gamma ⟨ɣ⟩ instead of Greek gamma ⟨γ⟩. But as you note, few fonts support it, and most IPA keyboards don’t use it, so usually the Greek letter is used.

(It would be nice if more fonts supported it though: sometimes chi is homoglyphic with ⟨x⟩, and using Latin chi avoids this problem…)

I don't know if the people at Unicode Inc. have changed their mind now, but that was not the original intention... The IPA Extensions block was there in 1991 in version 1.0, and it didn't contain IPA chi because you were supposed to use Greek chi. Most of Latin Extended-E (including U+AB53) was added recently in 2014 in version 7.0.

Oh, and I think the voiced velar fricative [ɣ] was encoded differently from Greek gamma because the IPA symbol is supposed to have a loop with a clear hole (or "eye" as typography people say), whereas Greek gamma very often has a tight loop with an inexistent eye.