Sound Change Quickie Thread

[Zaarin]

To switch the topic, if I may: how do glottalised consonants arise?

Cʔ and ʔC sequences are one possibility. Some Great Basin languages borrowed ejectives from their California neighbors and ended up with native ejectives. I understand implosives can arise from nasals, and I'm almost positive I've seen the change C^{2[/sup > Cʼ. In languages that contrast T Tʰ D, T > Tʼ to be more highly contrastive with the other two categories is a possibility. I've heard that some dialects of English have Tʼ as a word-final allophone of T, so apparently that's a possibility (in which case, those dialects would have a non-contrastive inventory of /T/ [T Tʰ Tʼ]; a Tlingit-speaker would be so confused...). Creaky voiced vowels can glottalize adjacent consonants.}

[Porphyrogenitos]

What are some possible outcomes of velarized consonants? Relatedly, if there was a "backing" influence on /p/ or /t/, would gaining a palatal articulation be a plausible outcome?

Context: I'm messing around with consonant harmony, with the sets /p t̪ k/ and /p̠ t̠ q/, and I want /k/ and /t̠/ to merge - /k/ will likely become /c/, I think.

[vokzhen]

Some Great Basin languages borrowed ejectives from their California neighbors and ended up with native ejectives.

Same appears to have happened in Southern Quechua from Aymara and Ossetian from Caucasian languages, most ejectives are borrowings but through either highly specific or irregular changes ended up in native words as well. From glances I've taken, they're clearly phonemic but also clearly in the minority - a Quechua dictionary I looked at they appear to be outnumbered about 10:1 by plain stops.

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

In languages that contrast T Tʰ D, T > Tʼ to be more highly contrastive with the other two categories is a possibility.

The examples I know of this - Southern Bantu, Eastern Armenian - are in contact with languages that already have T' Tʰ D (Khoisan and Caucasian languages, in those examples).

I've heard that some dialects of English have Tʼ as a word-final allophone of T, so apparently that's a possibility (in which case, those dialects would have a non-contrastive inventory of /T/ [T Tʰ Tʼ]; a Tlingit-speaker would be so confused...).

I'm reluctant to use this as an example just because there's (granted controversial) evidence Germanic *T originates from a glottalized segment, i.e. PIE *d [t'] or [ɗ~ʔd]. However, taking it as an innovation, it's similar to the the "plain" sets of both Javanese and Korean, which also have stiff/creaky voice, and could potentially harden into ejectivization.

Voiced stops can implosivize. In Austroasiatic, Vietnamese apparently had a merger of implosive and tenuis stops to implosives, and Munda languages tend to collapse coda stops to preglottalized voiced stops, but both started out with implosives. Some languages that have the labiovelar /gb/ allow debuccalization of the back contact, forming an implosive. I've heard gaining either implosives or ejectives from geminates is a possibility, but I have no examples on-hand. An extremely low tone that "bottoms out" into creakiness and then later shifts to consonant glottalization seems a possibility, though I'm not aware of any natlang precedents. Northern Ryukyuan languages apparently have them generally from ʔiCV and ʔuCV > ʔCV > C'V, but Yonaguni (Southern Ryukyuan) has them more generally from CV{high}CV > C2'V, provided both Cs are voiceless, with a concurrent CVCV > NCV when the second consonant is instead voiced. Siouan languages might be another place to look, where Mississippi Valley Siouan (including Dakotan) have ejectives but the other branches lack them, but I don't know how "mature" Siouan studies are and whether there's solid evidence of whether Dakotan and relatives gained them or the other branches lost them.

What are some possible outcomes of velarized consonants?

Athabascan tends to have /tʰ kʰ/ [tx kx], which merged to /kʰ/ [kx] in Jicarilla and I believe a few others. I wanna say I've seen something similar in Siouan languages, but I can't find where that might have been.

[Frislander]

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

The only one I can think of is Lushotseed, where the implosives arose from the denasalisation of glottalised resonants.

[Zaarin]

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

Sorry, it's just something I recall reading somewhere.

What are some possible outcomes of velarized consonants?

Athabascan tends to have /tʰ kʰ/ [tx kx], which merged to /kʰ/ [kx] in Jicarilla and I believe a few others. I wanna say I've seen something similar in Siouan languages, but I can't find where that might have been.

Lakhota has velarized release of aspirated consonants between back vowels, though as I understand it there are certain environments where they are marginally contrastive.

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

The only one I can think of is Lushotseed, where the implosives arose from the denasalisation of glottalised resonants.

I didn't realize there were any North American languages with implosives. In the other languages of the region sharing the "no nasal" feature like Nitinaht/Ditidaht, Snohomish/Sdohobish, Quileute, and Makah, the nasals became simple voiced stops. I'll have to read more about Lushootseed; it sounds like an interesting language.

[Frislander]

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

The only one I can think of is Lushotseed, where the implosives arose from the denasalisation of glottalised resonants.

I didn't realize there were any North American languages with implosives. In the other languages of the region sharing the "no nasal" feature like Nitinaht/Ditidaht, Snohomish/Sdohobish, Quileute, and Makah, the nasals became simple voiced stops. I'll have to read more about Lushootseed; it sounds like an interesting language.

Well for other native languages with implosives you also have Maidu and its relatives.

[k1234567890y]

Stage 1:

short: /a ə i u/
long: /a: e: i: o: u:/
diphthongs: /ai au əi əu/

from Stage 1 to Stage 2:

a > o
ə > e
a: > a
e: > ei
o: > ou
i: > ai
u: > au
ai > oi
əi > ei
əu > eu
au > ou

Stage 2:

/a e i o u/
/ai au ei eu oi ou/

[Zaarin]

I understand implosives can arise from nasals

Happen to have an example? The examples I know of is the reverse, implosives becoming nasal.

The only one I can think of is Lushotseed, where the implosives arose from the denasalisation of glottalised resonants.

I didn't realize there were any North American languages with implosives. In the other languages of the region sharing the "no nasal" feature like Nitinaht/Ditidaht, Snohomish/Sdohobish, Quileute, and Makah, the nasals became simple voiced stops. I'll have to read more about Lushootseed; it sounds like an interesting language.

Well for other native languages with implosives you also have Maidu and its relatives.

Cool. I didn't know there were implosives north of the Mayan languages; ejectives and glottalized sonorants are abundant in North America, but I was under the impression the implosives were non-existent--it seems instead they're simply rare.

[Nortaneous]

phpbb doesn't like form submission GET urls, but, here: http://bit.ly/2e7zTCp -- according to this PHOIBLE search thing, the only language north of Mexico with implosives is Maidu

(its Lushootseed data comes from UPSID, and UPSID thinks Lushootseed doesn't have implosives, but UPSID is very frequently wrong)

(edit: PHOIBLE's inventory of segments claims that there's such a thing as an implosive affricate, and four languages in Africa have prenasalized implosives)

[Chengjiang]

[mj] > [l]

I was thinking of this happening through an intermediate of [mlʲ] followed by simplification of the cluster and loss of the palatalization distinction, as [mj] > [mlʲ] happened in Russian, while former [lʲ] seems to have become simple [l] in Polish and General American simplified [lj] to [l]. (If you're wondering, this would most likely be alongside [pj] > [pl] and [bj] > [bl].)

Aside from that, I'm still interested in different ways languages have generated labialized velars. I'm also interested in generating lateral obstruents; I'm embarrassed to say the only case I know of that isn't fortition of [l] is Whorf's law ([t] > [tɬ] / _a).

[vokzhen]

[mj] > [l]

I'd buy it. My bigger concern would actually be whether mj>ml is attested outside of Slavic, or if it's kind of a freak occurrence.

Aside from that, I'm still interested in different ways languages have generated labialized velars.

Slavey has ts>tθ and then some dialects took it to either /p/ or /kʷ/, I assume when it fronted "too far" into [tf] territory and gained a labial element, then changed into the more common and more acoustically distinct /kʷ/ etc. Seri has /kmV/ > [kwṼ] which could be a source (it has genuine labialized velars/uvulars from deleted /o/). Initial sonorant fortition might generate it, e.g. j- w- > c- kʷ-. It might be possible to get them from labiovelars like /kp gb/. The vast majority of labialized velars seem to come from the obvious contact with /w/ or a deleted vowel, however.

I'm also interested in generating lateral obstruents; I'm embarrassed to say the only case I know of that isn't fortition of [l] is Whorf's law ([t] > [tɬ] / _a).

Once again fortition of /l/ seems the most common source (hl>ɬ, kl,tl>tɬ, ll>ɬ: etc). For a roundabout fortition, Khalkha Mongolian has l>ɮ, but devoices it word-initially, next to voiceless consonants, and occasionally just everywhere. s>ɬ occurs in some areas of Yue, e.g. Taishanese, is common in Central Tai, e.g. in Lungchow, and Khanty is posited to have *s *š > /ɬ/, plus in some dialects *l *ð > /ɬ/ and *ĺ > /ɬʲ/, followed by ɬ ɬʲ > t tʲ in others. Forest Nenets recently had a shift of r rʲ > ɬ ɬʲ. Getting it purely through borrowing doesn't seem horribly uncommon either, Mexican Spanish via Nahuatl and a bunch of languages in contact with Tibetan seem to have gotten it.

[----]

s > ɬ happens in an Athabaskan language (I can't remember which one, maybe Koyukon), along with ts > tɬ etc., and r > ɬ probably happened in Thao depending on how you want to reconstruct Proto Austronesian *R, e.g. *baqeRuh > faqɬu. r > ɬ definitely happened in Gedaged. From Proto-New Caledonia to Pwaamei we have ʈʰ > ɬ.

I'd also suggest looking at Bantu languages: several of them have lateral obstruents and Proto-Bantu had no (phonemic) laterals so you're guaranteed to get a few good correspondences there.

[Chengjiang]

I'd also suggest looking at Bantu languages: several of them have lateral obstruents and Proto-Bantu had no (phonemic) laterals so you're guaranteed to get a few good correspondences there.

I missed the bleeding obvious. Thanks, I'll look into that.

[----]

It was just a way of tricking you into doing the research that I was too lazy to do myself

[Frislander]

I'm back-deriving the proto-inventory for this language, and I have a phoneme /ʈʳ/ and I want it to split and caused a position-based alternation in the daughter language. Does changing it to /t/ word-initially and /t͡ʃ/ otherwise seem sensible?

[Knit Tie]

How could you derive a full set of contrasting dentals in a language that only has alveolar coronals?

[Zaarin]

How could you derive a full set of contrasting dentals in a language that only has alveolar coronals?

A few ideas:
t > t̪
c > t̺

t > t̺
c > t̪ʲ > t̪

t > t̺
θ > t̪

t > t̺
t(...)ʕ > t̪ˁ
t̪ˁ > t̪

t > t̪
ʈ > t̺

t > t̺
ʈ > t̪

t > t̪
ʦ, ʧ > t̺

t > t̪
k > t̺
q > k OR p > kʷ > k

Also, as with ejectives, this is a good candidate for loanwords if it's in intimate contact with another language that already has this contrast.

[k1234567890y]

is the following sound changes plausible?

Stage 1:

short: /ɛ ɑ (ə) i u/
long: /ɛ: (ɑ:)/

diphthong: /ɛi ɛu ɑi ɑu/ and probably /ɛ:i ɛ:u/

from Stage 1 to Stage 2:

əi əu > i u
əH > a: > a(H is an laryngeal consonant)
ɛ ɑ > a > ə
ɛ: ɛH ɑH > a: > a(H is an laryngeal consonant)
ɛi ɑi > ai > əi
ɛu ɑu > au > əu
ɛɟ ɑɟ > aʝ > ai
ɛg ɑg > aɣ > au
iH uH > i: u:
iɟ > iʝ > i:
ig > iɣ > i:
uɟ > uʝ > uj > u:(or i:?)
ug > uɣ > u:

Stage 2:

short: /a ə i u/
long: /a: i: u:/
diphthongs: /ai au əi əu/

the time span between different stages are uncertain

[Tropylium]

What are some possible outcomes of velarized consonants?

ɫ > w or ɫ > ɣ are common as dirt, but you probably knew that.

Velarization > pharyngealization > voicing, as in Arabic & co., would probably be reasonable. Proto-Oceanic *pʷ may also have been simply [pˠ] (as it is in e.g. Marshallese, IIRC); it often shifts to /kʷ/, or to plain /p/ (while unvelarized *p > *ɸ > f, h, β, etc.).

If there was a "backing" influence on /p/ or /t/, would gaining a palatal articulation be a plausible outcome?

Probably not. Depends on what this "backing" is being exactly — /u/ can sometimes palatalize things (Japonic, Bantu), but things like /o/ definitely cannot.

[k1234567890y]

Probably not. Depends on what this "backing" is being exactly — /u/ can sometimes palatalize things (Japonic, Bantu), but things like /o/ definitely cannot.

it seems that cross-linguistically, high vowels /i/ and /u/ actually often behave in a similar manner?

[Chengjiang]

Probably not. Depends on what this "backing" is being exactly — /u/ can sometimes palatalize things (Japonic, Bantu), but things like /o/ definitely cannot.

I thought the "palatalization" before /u/ was less actual palatalization and more (af)frication before a [+close] segment.

[Knit Tie]

ɫ > w or ɫ > ɣ are common as dirt, but you probably knew that.

What about the reverse?

[----]

I don't know if either is attested, but we do have velars becoming alveolar liquids in some instances. Pohnpeian and a couple other Micronesian languages have the correspondence POc *ⁿg > PMic *x > r/_a, with *x probably being ɣ based on other correspondences. (it's not attested as a velar in any of the modern languages except Kosraean which reflects it as /k/) So we have correspondences like
PMic *xawu > Pohnpeian reew ; POc *waⁿga > Pohnpeian waar , etc. r > l is not an odd change at all so it's not impossible for ɣ > l to happen, by my count.

[Sexendèƚo]

What peculiarities might we see in a language which does not have vowel harmony, but whose ancestor did?
I'm working on my conlang's vowel system, and I would like a situation where the ancestor had front-back harmony, but this system collapsed. I want it to still show up vestigially in the language though.

The proto-lang has a simple inventory, with 5 front-back allopohonic pairs: /æ~ɒ/,/e~ɤ/,/i~ɯ/,/o~ø/,/u~y/. Frontness/backness is a word-level feature.
Here are my ideas so far:
1. Vowels in affixes change differently depending on the allophone, leading to declensions/conjugations. Something like this:
/-usseː/
[+front] [-ysse]→[isse]
[+back] [-ussɤ]→[usse]
2. Front/back pairs which are long dipthongize differently.
/noː-/
[+front] [nøː-]→[neʉ̯-]
[+back] [noː-]→[nou̯-]
3. Front triggers palatalization, but not back.
[løkki]→[locci]
[lokkɯ]→[lokki]

Do these seem plausible? Any other ideas?

[Zaarin]

I have a rather weird one I want to ask about. s/k/_# Does that seem too far fetched?