Sound Change Quickie Thread

[sangi39]

Is /θ ð/ > /w/ possible (with or without intermediate steps)?

With enough intermediate steps, everything is possible.
/θ ð/ > /f v/ is attested, and you could take it from there.

I'd assume you could have [ð] > [ɣ] as in Gaelic as well, and then something like [ɣ] > ([ɰ]) > [w] (conditional in Old English > Modern English).

Does anyone know what actually happened in Faroese in relation to <ð>? I'm assuming that it's etymologically determined, and thus all instances of <ð> represent instances of [ð] in Old West Norse, but was this sound dropped, with glide insertion to break vowel hiatus occurring later, or did it become [j] ~ [w] ~ [v] ~ [0] as a kind of split. <g> seems to be used in the same way, so I wonder if there was a kind of merger between Old West Norse [ð] and [g] at some point, at least between vowels, with the resulting phone shifting to glides.

[jmcd]

Directly would be implausible if the resultant phonology doesn't contain few phonemes. Indirectly via [ɣ] or [ɹ] for example, yes.

[Pole, the]

Is /θ ð/ > /w/ possible (with or without intermediate steps)?

I guess. I once had /θ ð/ → /f w/ in one of my conlangs.

[Chengjiang]

Is /θ ð/ > /w/ possible (with or without intermediate steps)?

I take it you mean unconditionally? At any rate, I don't see why not. [ð] can easily shift to [ɣ] or [ɰ] and thence to [w], especially since [ð], as a dental consonant, is often at least slightly velarized. It can also shift to [v] or [ʋ] and get to [w] that way. Really, any voiced fricative that isn't palatal(ized) or sibilant can go to [w] with little or no intermediary. [θ] to [w] is a little bit more unusual, but assuming you don't want it to happen by [θ] merging into [ð] first, you can still get it via, say, [f] with subsequent lenition to [w].

[Tropylium]

†rd

What does the † mean?

"Attested but only in an extinct variety" (in this case, Old Persian).

Which may seem sort of superfluous since we can call any prior stage of linguistic development technically "extinct", but it can be contrasted with "still attested in the same form from a different descendant".

(Of course, we could also get into further increasingly philosophical debate on if non-Persid Iranian languages that still have /rd/, say Balochi, would count…)

[Nannalu]

Is there any way I could create /ʕ/ and /Cˤ/ from an SAE inventory?

[Pole, the]

r → ʕ
Cʕ → Cˤ

[Nortaneous]

Vr > V_?\ happened in Chemnitz German, and once you have pharyngealization it's not hard to push it around between vowels and consonants (see: Chechen vs. Ingush)

[Travis B.]

One set of changes I had in one of my languages is:

ʁ > ʕ / _{a ɑ}
χ > ħ / _{a ɑ}

(After which the conditioning environment was obscured.)

Note that this can also happen unconditionally:

ʁ > ʕ
χ > ħ

[Pogostick Man]

/g/ became pharyngeal in one European language. I want to say it was Iberian in origin but the name escapes me right now.

[Atrulfal]

/g/ became pharyngeal in one European language. I want to say it was Iberian in origin but the name escapes me right now.

Galician ?

[jmcd]

/g/ became pharyngeal in one European language. I want to say it was Iberian in origin but the name escapes me right now.

You are likely thinking of Galician, where [ħ] can be an allophone of /g/ presumably intervocalically.

[vokzhen]

/g/ became pharyngeal in one European language. I want to say it was Iberian in origin but the name escapes me right now.

You are likely thinking of Galician, where [ħ] can be an allophone of /g/ presumably intervocalically.

Yar, the intervocal [ɣ] back and devoiced. An intermediate happen in Ukrainian, where in some words Russian, Polish, and Serbo-Croatian have /g/, Czech, Slovak, and Upper Sorbian have /ɦ/ or /h/, and Ukrainian has [ʕ].

[WeepingElf]

Do you think these two make sense?

ts > t'
q > k'

I know that the reverse has happened in some Semitic languages, but are there examples for this direction, in a language that previously doesn't have ejectives?

[TinyMusic]

Is this attested? If not, is it possible?:
θ -> kʲ
ð -> gʲ

[StrangerCoug]

Is this attested? If not, is it possible?:
θ -> kʲ
ð -> gʲ

The way I see this happening is to fortite /θ ð/ to /t d/ and go from there, possibly after getting ideas from Hawaiian (which had t k).

[vokzhen]

Do you think these two make sense?

ts > t'
q > k'

I know that the reverse has happened in some Semitic languages, but are there examples for this direction, in a language that previously doesn't have ejectives?

The only way I know to gain ejectives for sure like that is the aspirate/plain/voiced>aspirate/ejective/voiced found in Eastern Armenian and Southern Bantu, in contact with Caucasian and Khoisan that also have aspirate/ejective/voiced sets. However, the Korean fortis series, the plain series in Javanese, and the plain series in Zaiwa/Atsi and Langsu/Maru (sister Burmish languages) have glottalization on the following vowel, so possibly via that. Either way I don't know how you'd get /q ts/ to yield /k' t'/, though, without mergers or other shifts involving /k' t'/ from /k t/.

Is this attested? If not, is it possible?:
θ -> kʲ
ð -> gʲ

I might be able to see it as θ ð > ç ʝ > kʲ gʲ but I'd expect other fricative fortitions too.

[Zaarin]

Do you think these two make sense?

ts > t'
q > k'

I know that the reverse has happened in some Semitic languages, but are there examples for this direction, in a language that previously doesn't have ejectives?

Note that what happened in said Semitic languages is k' > kˁ > q and t' > tˁ > ʦ.

On which note, I have a language that has ejectives in which I plan to use the Aramaic/Arabic ejective-to-pharyngealized shift. However, said language already has /q q'/ and I don't want to merge historical /k'/ and /q/ (i.e., I want to change /q/ into something else first). One possibility, of course, is q > ʔ, but what are some other ways I can get rid of /q/ prior to the merger? I don't want q > k because I did that in a sister language. I've also considered q > χ > ħ, but I'd like to consider some other possibilities.

[opipik]

q > X* > x
x > ç / _i**

----

*uvular
**optional

[Porphyrogenitos]

On which note, I have a language that has ejectives in which I plan to use the Aramaic/Arabic ejective-to-pharyngealized shift. However, said language already has /q q'/ and I don't want to merge historical /k'/ and /q/ (i.e., I want to change /q/ into something else first). One possibility, of course, is q > ʔ, but what are some other ways I can get rid of /q/ prior to the merger? I don't want q > k because I did that in a sister language. I've also considered q > χ > ħ, but I'd like to consider some other possibilities.

Well, there's /ʀ̥/, or you could just leave it at /χ/. I'd suggest /q͡χ/, but there are apparently no natural languages that contrast /q͡χ/ with /q/, IIRC.

[Zaarin]

q > X* > x
x > ç / _i**

The language already has χ x > ħ, like Phoenician. Like I said, I'm considering that option; none of the other languages in the family lenite /q/--even the one that has post-vocalic lenition of the other plosives--so it would be different.

EDIT: Don't some dialects of Arabic realize /q/ as [g]? Does anyone know the philology of that?

[vokzhen]

EDIT: Don't some dialects of Arabic realize /q/ as [g]? Does anyone know the philology of that?

It fills in for the missing /g/, since most dialects have /t k q/ and /b d/, a step towards leveling the system. Not sure if you could pull it off already having /g/. There's dialects that have the emphatic series as no aspiration and the voiceless series as lightly aspirated; off the top of my head I'm not sure how they relate to the dialects that have q>g but I imagine they overlap. I've also heard Classical Arabic actually had [ɢ], in which case a reflex of /g/ might just be fronting, but I'm not sure I buy that (afaik that's based on /q/ being classified by Classical Arabic grammarians along with /b d/ rather than /t k/, but a voiceless, unaspirated could have been thrown in with either).

By far the two most common changes are q>χ and q>ʔ. A "satemization" chain of q>k>tS happens too, I'm not sure how commonly but most of Mayan clearly has it, as less ambiguous, non-PIE example. ʡ happens occasionally rather than becoming a fricative (Northern Haida, and afaik that's what happened in Amis, and Somali's qʡ).

[Zaarin]

EDIT: Don't some dialects of Arabic realize /q/ as [g]? Does anyone know the philology of that?

It fills in for the missing /g/, since most dialects have /t k q/ and /b d/, a step towards leveling the system. Not sure if you could pull it off already having /g/. There's dialects that have the emphatic series as no aspiration and the voiceless series as lightly aspirated; off the top of my head I'm not sure how they relate to the dialects that have q>g but I imagine they overlap. I've also heard Classical Arabic actually had [ɢ], in which case a reflex of /g/ might just be fronting, but I'm not sure I buy that (afaik that's based on /q/ being classified by Classical Arabic grammarians along with /b d/ rather than /t k/, but a voiceless, unaspirated could have been thrown in with either).

By far the two most common changes are q>χ and q>ʔ. A "satemization" chain of q>k>tS happens too, I'm not sure how commonly but most of Mayan clearly has it, as less ambiguous, non-PIE example. ʡ happens occasionally rather than becoming a fricative (Northern Haida, and afaik that's what happened in Amis, and Somali's qʡ).

Hmm, that does actually give me an idea: g > ɣ > ʕ (already paralleling the x > ħ), then q > g. I'll definitely add that to the list of possibilities at least.

[KathTheDragon]

Is *V *Vh > Ø V plausible? V can be any vowel, long or short.

[Chengjiang]

Is there any way I could create /ʕ/ and /Cˤ/ from an SAE inventory?

To add to the posts suggesting deriving them from rhotics, dorsals, and clusters thereof, I'd also like to note that dental consonants are often at least slightly pharyngealized naturally, especially [l]. (Hence "clear L" versus "dark L".) You can easily have a contrast between two types of coronal consonant turn into a pharyngealized/clear contrast, e.g. lamino-dental /ts dz s/ versus apico-alveolar /ts dz s/ turning into /tsˤ dzˤ sˤ/ versus /ts dz s/. And since I brought up clear and dark L, laterals are another good source of pharyngeal consonants or pharyngealization, e.g. [VlC] > [VlˤC] > your pick of [VʕC], [VˤC], or [VCˤ].

Addendum: Retracted vowels, a.k.a. open back vowels, are also typically somewhat pharyngealized to begin with, and so consonants or other vowels in their vicinity may become pharyngealized. You could have, say, [C] > [Cˤ] when the following vowel is one of [ɑ ɒ ʌ ɔ] and then do some vowel shifts to make the pharyngealization contrastive. You could also have diphthongs with an opening glide (e.g. non-rhotic German's diphthongs ending in non-syllabic [ɐ]) become pharyngealized vowels.