Sound Change Quickie Thread

[احمکي ارش-ھجن]

How can I go from n > 0 intervocalically? Preferably in as little steps as possible?

Would n > nd > d > t > ? > 0 work?

[Pole, the]

Maybe [n] → [ɹ̃] → ∅.

[jmcd]

Follow Portuguese

[Vijay]

does anyone here know any examples of a palatal or alveolopalatal turning into a velar?

Didn't Egyptian Arabic turn /dʒ/ into /g/?

Apparently, in Classical Arabic, it was [ɡʲ] or [ɟ], not [dʒ]. But I know this change happened in Kalderash Vlax Romani, for example, so e.g. 'your' is kiro in Kalderash from earlier čiro, and 'hat' is stagi from earlier stadji.

[vokzhen]

Didn't Egyptian Arabic turn /dʒ/ into /g/?

Classical Arabic had /gʲ~ɟ/ that fronted in most dialects to [dʒ~ʒ~j]. It seems more likely Egyptian Arabic backed Classical Arabic /ɟ/ without an intermediate. Or perhaps it simply adopted [g] for /ɟ/ due to Coptic or Greek influence.

How can I go from n > 0 intervocalically? Preferably in as little steps as possible?

Would n > nd > d > t > ? > 0 work?

Fastest I know of is probably Basque, which took n>h, and intervocal h>zero is extremely common. Of course n>h is a n extremely unusual change in the first place. n > r > zero could probably work. Another way might be via [ɨ̃] (or, as Pole said [ɹ̃]), but I'm not sure you'd realistically lose the nasalization without an intermediary of a (probably coalesced) nasal vowel, e.g. /ane/ > [aɨ̃e] > [ɛ̃ɛ̃] > [ɛɛ].

[Vijay]

Fastest I know of is probably Basque, which took n>h

It did?!

[vokzhen]

Well, intervocally yes, except it looks like Vni# > Vɲ and ini > ihi or i.

[احمکي ارش-ھجن]

Eh, I just wanted to make a Romance language where the indefinate articles were uo and ua, but I don't wanna make Portuguese.

[KathTheDragon]

I suppose you could possibly do n>m in the context of u, and then lenite m intervocalically, ending up with w, and then have aphaeresis of the initial vowel (perhaps generally, or specifically uw>u/#_)

[Vijay]

Do you even have to get rid of the [w]?

This reminds me of Romani. In the Central and Eastern Indo-Aryan languages, [m] changed to [ʋ] with nasalization on the preceding vowel. But in Romani, the nasalized vowels merged with the oral ones. For example, 'village' in Sanskrit (for example) was grāma, in Hindi it's [gãʋ], and in Romani it's [gaʋ].

[KathTheDragon]

Ahzoh wasn't clear on how those were pronounced.

[gach]

Wouldn't just plain deletion n > ∅ in fast speech or unstressed environments work fine? I don't see any particular need for intermediate stages there, especially in the context of common grammatical words like articles.

[Atrulfal]

I wanted to go from

/m n ɲ/
/p b t d k g/
/β f s z ʃ ʒ/
/ts dz tʃ dʒ/
/l ʎ r ɾ/

to

/m n ŋ/
/p pʰ t tʰ k kʰ/
/ts tʃ/
/f v s ʃ x h/
/j l ɾ/

And to achieve that I came up with followings changes

n > h/V_V
h > ŋ
ʒ > ʃ
dz dʒ > ts tʃ
β > v
r > x
z > r
ʎ ɲ > jl jn
x > h
r > x
p b t d k g > pʰ p tʰ t kʰ k

Help ?

[sangi39]

n > h/V_V - Discussed above in relation to a question from Ahzoh, seems plausible enough
h > ŋ - I'm still not sure on this from the last time you asked. Maybe?
ʒ > ʃ - Old Spanish > Modern Spanish, if I remember rightly
dz dʒ > ts tʃ - The first happened, again, from Old Spanish to Modern Spanish, the second doesn't seem implausible at all as a result.
β > v - Again, easy transition, makes sense
r > x - Some Portuguese dialects have this, don't they?
z > r - Again, non-Eastern Germanic and Latin
ʎ ɲ > jl jn - Early Old French had something similar, e.g. [zj] > [zʲ] > [jzʲ] > [jz] in Latin mansiōnātam > Old French maisniée, so I wouldn't call it unlikely.
x > h - Doesn't seem at all unlikely
r > x - Listed twice
p b t d k g > pʰ p tʰ t kʰ k - A fair number of languages, I think. First one that comes to mind is Icelandic.

[Vijay]

n > h/V_V - Discussed above in relation to a question from Ahzoh, seems plausible enough
h > ŋ - I'm still not sure on this from the last time you asked. Maybe?

Wouldn't it be easier to just go from n directly to ŋ?

[Atrulfal]

r > x - Listed twice

This /r/ is the new one that came from z > r.

n > h/V_V - Discussed above in relation to a question from Ahzoh, seems plausible enough
h > ŋ - I'm still not sure on this from the last time you asked. Maybe?

Happened in Nyole. There it went like this n > h̃ > ŋ .

ʎ ɲ > jl jn

Happened in Basque.

Wouldn't it be easier to just go from n directly to ŋ?

How ?

[Ser]

n > ŋ is attested in Andalusian/Caribbean/Central American Spanish, but only in word-final position. Ganan [ˈga.naŋ ~ ɣ̞-].

[Vijay]

Apparently, there's a variety of Romanian that has n > ŋ spoken in Țara Moților in Transylvania.

[Chengjiang]

Wouldn't just plain deletion n > ∅ in fast speech or unstressed environments work fine? I don't see any particular need for intermediate stages there, especially in the context of common grammatical words like articles.

This. English had [n] > [0] in many grammatical particles, retaining only an as a variant of a (and, in Early Modern English, mine and thine of my and thy) used before vowel-initial words.

[Pogostick Man]

I'm looking to generate a contrast between voiceless, slack-voiced, and voiced consonants (and possibly even stiff-voiced ones as well). How might I go about developing slack and stiff voicing?

[Vijay]

Well, I don't know how much this helps, but I found this. On p. 109-110, they appear to be saying that in Khmer, it's possible that voiced and voiceless initials became stiff and slack-voiced. I'm not sure whether there are any languages out there that have a three- or four-way voicing contrast of the kind you're trying to generate, though...

[ivazaéun]

Is ɬ θ attested? Would it require intermediary steps?

Based on that, is it reasonable to generate a dental/alveolar split like this?
tɬ,tl tθ t̪

[Zaarin]

Is ɬ θ attested? Would it require intermediary steps?

Yes, directly in Proto-Muskogean > Creek.

[Vijay]

Creek has [θ] and not [ɬ]?

EDIT: Woods Cree does have [ð], which apparently comes from Proto-Cree *l.

[Zaarin]

Creek has [θ] and not [ɬ]?

EDIT: Woods Cree does have [ð], which apparently comes from Proto-Cree *l.

Maybe it's not Creek; maybe it's Choctaw-Chickasaw. At least one Muskogean language has ɬ > θ, anyway.