Sound Change Quickie Thread

[bbbosborne]

I would think you could get P' from sP extremely simply, just by sP > hP > PP > P'

I mean, I think hP > PP is plausible. It could go hP > ʔP > PP, and that would be plausible for sure, right?

yea, PP is literally ʔP.

[dhok]

*ʔP > hP is attested in most of Algonquian (Blackfoot, Arapaho, Cheyenne, Cree and Menominee excepted); *hP > PP is attested in Ojibwe, and *hp > ʔP in Shawnee.

[linguistcat]

Deriving a language from Old Japanese, which was pretty strictly CV with some word-initial V syllables. After some vowel loss, I plan to have various sound changes happening to the various consonant combos that result from this. I can deal easily with nasal+stop and I might just make 2 of the same consonant a long consonant, which would be before it happened in Japanese AFAIK. Having trouble deciding what to do with two different stops, stop+s or s+prenasalized stop sequences. My first instinct for this language is to say that the first stop of two becomes a fricative, but I like that best when it's two plain stops and would like to do something more interesting with sequences involving prenalized stops (and z, which was also prenasalized).

I'd like this language to sound "cat like" but I also know that what that means varies a little depending on people's own experiences. But for me that would suggest more nasals, more fricatives/trills and possibly making it tonal instead of relying on pitch accent.

[mèþru]

Let N = nasal, S = stop, F = fricative, and P = prenasalisation

PS/N/_PS
SP/N/_S
S/F/(P)_S

[linguistcat]

Let N = nasal, S = stop, F = fricative, and P = prenasalisation

PS/N/_PS
PS/N/_S
S/F/(P)_S

These are actually pretty close to what I came up with. On the last sound change, what do you mean by putting (P) in the condition like that? I want to make sure I understand what you were going for.

[mèþru]

Just making it clear that it doesn't matter whether the first stop is prenasalised or not for that rule

[linguistcat]

Ah, I get what you mean then. I might play around a bit, especially since I still have other combos to decide.

[missals]

I have recently speculated a bit about a Romlang in which the Latin vowels turn out as:

Latin > Romlang

i i: > e i
e e: > e i
a a: > a a
o o: > o u
u u: > o u

Now, I am certain this is possible if it evolved directly from the Latin vowel system, since the length distinction would keep /e:/ and /i/ (etc) from merging before the quality distinction collapsed.

However, I would want to derive it from Proto-Romance, where the distinction between the vowels was primarily by quality, i.e.

Latin > PR > Romlang

i i: > ɪ i(:) > e i
e e: > ɛ e(:) > e i
a a: > a a > a a
o o: > ɔ o(:) > o u
u u: > ʊ u(:) > o u

Without relying on any possible remaining length distinction, is this possible? Has anything like this happened before?

I suspect that maybe the "laxness" of the historical short vowels and the "tenseness" of historical long vowels could result in /ɪ ɛ ɔ ʊ/ being a bit more centralized and /i e o u/ being a bit more peripheral, thus avoiding a merger between the qualitative pairs, and encouraging a merger between the quantitative pairs.

[Knit Tie]

What do you think about this sound chanɡe, where pharynɡealisation/uvilarisation on consonants becomes a vowel quality distinction?

i u ɛ → e o æ/ Cˤ_; _Cˤ
æ ɑ →a
e ɛ →e

This results in a classic 5-vowel system.

Also, how's this diachronic shift chain, assuminɡ that /ʕ/ exists in the lanɡuaɡe already?

ɡ → ʁ →ʕ →ŋ

[dhok]

What do you think about this sound chanɡe, where pharynɡealisation/uvilarisation on consonants becomes a vowel quality distinction?

i u ɛ → e o æ/ Cˤ_; _Cˤ
æ ɑ →a
e ɛ →e

This results in a classic 5-vowel system.

Also, how's this diachronic shift chain, assuminɡ that /ʕ/ exists in the lanɡuaɡe already?

ɡ → ʁ →ʕ →ŋ

I'm pretty sure exactly that sort of vowel shift has happened in at least one dialect of Arabic--Moroccan, maybe?

*g > ŋ unconditionally has occurred in Čiwere, a Siouan language. *ʕ > ŋ may have occurred in some dialects of Hebrew, but may be an L1-influenced substitution; ʕ seems not to fortite much. (But then again, it's so rare that there just aren't many known changes.)

[Zaarin]

What do you think about this sound chanɡe, where pharynɡealisation/uvilarisation on consonants becomes a vowel quality distinction?

i u ɛ → e o æ/ Cˤ_; _Cˤ
æ ɑ →a
e ɛ →e

This results in a classic 5-vowel system.

Also, how's this diachronic shift chain, assuminɡ that /ʕ/ exists in the lanɡuaɡe already?

ɡ → ʁ →ʕ →ŋ

The vowels look fine. The second shift I'd either go g > ŋ as Dhok suggested or add an extra step: ɡ > ʁ > ʕ > ʔ > ŋ.

[Pabappa]

ʕ →ŋ

I may be the source of this myth, to be honest .... I read something long ago about Hebrew and repeated it here, and it got bundled into the Index Diachronica. Though in my defense, even there, it says that it may have been a substitution rather than a true sound change .... basically what dhok says.

I dont think unconditional /ʕ/ > /ŋ/ is a reasonable sound change. I dont think /ʔ/ > /ŋ/ works either. On the other hand, I'm sure /g/ > /ŋ/ is fine, perhaps with a middle stage of /ŋg/ at least between vowels. I would just use /g/ > /ŋ/ unless you have a specific reason to also bump off some other sounds.

[Knit Tie]

I thought that /ʕ/ > /ŋ/ was a better sound change than /ʔ/ > /ŋ/ due to rhinoglottophilia? In any case, /ʕ/ is, as dhok said, very rare, so I think I can just make stuff up here with little consequence. I do want to turn /g/ into the velar nasal via /ʕ/ to merge it with preexistent /ʕ/, though.

[Zaarin]

I dont think /ʔ/ > /ŋ/ works either.

I'm reasonably sure it's attested, though, and it's pretty straightforward rhinoglottophilia.

[Ælfwine]

I have recently speculated a bit about a Romlang in which the Latin vowels turn out as:

Latin > Romlang

i i: > e i
e e: > e i
a a: > a a
o o: > o u
u u: > o u

Now, I am certain this is possible if it evolved directly from the Latin vowel system, since the length distinction would keep /e:/ and /i/ (etc) from merging before the quality distinction collapsed.

However, I would want to derive it from Proto-Romance, where the distinction between the vowels was primarily by quality, i.e.

Latin > PR > Romlang

i i: > ɪ i(:) > e i
e e: > ɛ e(:) > e i
a a: > a a > a a
o o: > ɔ o(:) > o u
u u: > ʊ u(:) > o u

Without relying on any possible remaining length distinction, is this possible? Has anything like this happened before?

I suspect that maybe the "laxness" of the historical short vowels and the "tenseness" of historical long vowels could result in /ɪ ɛ ɔ ʊ/ being a bit more centralized and /i e o u/ being a bit more peripheral, thus avoiding a merger between the qualitative pairs, and encouraging a merger between the quantitative pairs.

I don't think anything like that has happened in the history of Romance. Though its not impossible, I think in Taravo you have a distinction preserved between short /i/ and long /eː/.

Apparently, in the northwestern Balkan region, the Romance vowel system had a tendency to confuse long /eː/ with /iː/ (but not I > E). This has been attributed to Greek influence, although it was only a passing phase. In analogy, you could have something like /oː/ > /uː/ happen as well.

[missals]

I don't think anything like that has happened in the history of Romance. Though its not impossible, I think in Taravo you have a distinction preserved between short /i/ and long /eː/.

Apparently, in the northwestern Balkan region, the Romance vowel system had a tendency to confuse long /eː/ with /iː/ (but not I > E). This has been attributed to Greek influence, although it was only a passing phase. In analogy, you could have something like /oː/ > /uː/ happen as well.

Thanks for the info! I knew that nothing quite like this had happened in Romance, but I was curious if it was plausible typologically speaking, I guess.

Another thing I've been thinking of (inspired somewhat by English vowel allophony):

We start out with five vowels, /a e i o u/

o u > ə ɨ, except before [l], which vocalizes/drops, resulting in:

al > aɒ̯ > ɒ
el > eo̯
il > iu̯
ol > oo̯ > o
ul > uu̯ > u

(This creates fun alternations such as tamo/tamol > tamə/tamo)

But if original /u/ fronts, it seems appropriate for original [u̯] to front, and I am thinking of having an original [u̯] in a series of dipthongs like:

[au̯ eu̯ iu̯ ou̯]

Which would then become:

[aɨ̯ eɨ̯ iɨ̯ oɨ̯]

These don't seem very stable. It seems like they might be prone to further fronting, like to [ai̯ ei̯ i oi̯], which I guess works, but does anyone have any other ideas for what they could become? Especially since it's not syllabic, I feel that the [ɨ̯] offglide is likely to become a consonant. But what consonant is [ɨ̯] likely to become?

[Zaarin]

But what consonant is [ɨ̯] likely to become?

/ɰ/ or /j/ seem the likeliest to me.

[bbbosborne]

is /pʃ/ > /ʙ/ plausible?

[Xwtek]

is /pʃ/ > /ʙ/ plausible?

I don't think so, /ʙ/ usually is a result of lenition of /b/, like /r/ is a result of lenition of /d/ or /z/ (i.e. Proto Germanic > North and West Germanic)

[Xwtek]

But what consonant is [ɨ̯] likely to become?

For me, it is /ɹ̪/.