Sound Change Quickie Thread

[XYZ]

So let’s say I have the following system of stops and fricatives within a larger inventory (simplified slightly for clarity):
/p t k/
/ᵐb ⁿd ᵑɡ/
/s ⁿz/
/s/ and /ⁿz/ have the allophones [ʃ] and [ⁿʒ] before /i/.

Following pervasive apocope, many prenasalized consonants (which before only appeared in medial positions), can now be found in coda position. In this position I want the prenasalized stops to change into their simple nasal equivalents, much like what one may see in many of the languages of Torba Province in Vanuatu. For example:
/aᵐba/ > /aᵐb/ > /am/
/aⁿda/ > /aⁿd/ > /an/
/aᵑɡa/ > /aᵑɡ/ > /aŋ/
The prenasalized consonants are retained medially unchanged.

What I can’t figure out is an equivalent sound change that would work for /ⁿz/ and its allophone [ⁿʒ] (which I guess would become phonemic following the apocope). Maybe because I’m not too skilled at prenasalizing these sounds (and end up producing something that sounds like fully nasalized fricatives), but I think a possible solution is just to denasalize it completely and leave us with /z/ and /ʒ/ in all positions, but I thought it was kind of boring. Another idea I had was to turn those two into some sort of nasalized approximants word-finally to somewhat mirror changes like /aᵐba/ > /am/ even if not 100% parallel, but even that’s giving me hold-ups. I feel like I should just throw out the phoneme but I don’t want to directly “steal” from the languages I’m inspired by. This might be a bit long for a quicky but any suggestions or thoughts. would be appreciated.

[Pabappa]

looks like Old Japanese. id suggest one of two things:
1) /nz/ > /z/ as you propose. this could come from allophonic nasalization of the preceding vowel that then disappears. and this may even be what happened in Modern Japanese.
2) /nz/ > /ndz/ > /nd/, then conditionally > /n/ following the pattern of the original /nd/. I think /z/ has disappeared in Sanskrit in similar positions but Im not sure.

A nasalized approximant could theoretically shift to /n/, but Im not sure I'd believe that the approximant stage would be stable enough to have a single outcome in all positions.

[Nortaneous]

What I can’t figure out is an equivalent sound change that would work for /ⁿz/ and its allophone [ⁿʒ] (which I guess would become phonemic following the apocope).

Depends on the inventory. In some Papuan languages, /s ⁿz/ are actually [s~ʃ ⁿdʒ] and pattern as palatals. So you could back /ⁿz/ to /ⁿʒ/ and have it nasalize to /ɲ/.

[Pabappa]

/v/ > /b/ happened in Spanish. /f/ > /p/ is sometimes claimed for early Semitic, but i suspect this sound change is very rare, so perhaps first /f/ > /v/, then /v/ > /b/. Youd still have /p/ from original /p/.

That was quick!

Okay, that might actually work. I have also considered debuccalizing /f/ (potential Basque adstrate?) Though despite its rarity, I don't see much issue with /f/ > /p/ if my substrate in question simply pronounces all /f/ as /p/, although for that I'd like a second opinion.

I meant to add that I suspect most /f/ > /h/ is really /ɸ/ > /h/. For sure it was in Japanese and Celtic, and it seems likely also in Polynesian. Spanish has true /f/ > /h/ in initial position, though.

[Tropylium]

Is
i a u>e o o/_C*V
where V is an identical vowel.
realistic

Identical to the input or identical to the output? Vowel assimilation will make some sense almost always.* Dissimilation might be a harder sell in this case.

If it's the former, I'd go a bit further even: if you have a change like i…e > e…e, I'd also expect to see some not-quite-all-the-way assimilations like i…a > e…a.

* For some reason this does not work very well for consonants. Assimilation will still always be plausible in clusters, but in the case of consonants separated by a vowel, then dissimilation is maybe more common, both for places and manners of articulation. I guess nasality is the main exception, and then that could be due to transmission by the vowel: a change like "ban > man" is possibly in more detail [ban] > [bãn] > [mãn].

[Frislander]

What I can’t figure out is an equivalent sound change that would work for /ⁿz/ and its allophone [ⁿʒ] (which I guess would become phonemic following the apocope). Maybe because I’m not too skilled at prenasalizing these sounds (and end up producing something that sounds like fully nasalized fricatives), but I think a possible solution is just to denasalize it completely and leave us with /z/ and /ʒ/ in all positions, but I thought it was kind of boring. Another idea I had was to turn those two into some sort of nasalized approximants word-finally to somewhat mirror changes like /aᵐba/ > /am/ even if not 100% parallel, but even that’s giving me hold-ups. I feel like I should just throw out the phoneme but I don’t want to directly “steal” from the languages I’m inspired by. This might be a bit long for a quicky but any suggestions or thoughts. would be appreciated.

I've definitely noticed a slight tendency diachronically for sibilants to delete preceding nasals, particularly in Indo-European but also outside of it, so I think simply denasalisation to /z/ is perfectly legitimate.

Is
i a u>e o o/_C*V
where V is an identical vowel.
realistic

I echo Tropylium here, and additionally I would say this seems a little odd a formulation as well. Is this supposed to be with subsequent apocope to obscure the conditioning?

[Ælfwine]

/v/ > /b/ happened in Spanish. /f/ > /p/ is sometimes claimed for early Semitic, but i suspect this sound change is very rare, so perhaps first /f/ > /v/, then /v/ > /b/. Youd still have /p/ from original /p/.

That was quick!

Okay, that might actually work. I have also considered debuccalizing /f/ (potential Basque adstrate?) Though despite its rarity, I don't see much issue with /f/ > /p/ if my substrate in question simply pronounces all /f/ as /p/, although for that I'd like a second opinion.

I meant to add that I suspect most /f/ > /h/ is really /ɸ/ > /h/. For sure it was in Japanese and Celtic, and it seems likely also in Polynesian. Spanish has true /f/ > /h/ in initial position, though.

That was also my thinking. I could attribute to a Basque substrate as much as an Innu substrate, much like in Spanish, as Basque also supposedly lacked this phoneme. (And IIRC Basque whalers/traders were active in the region.)

[Zaarin]

Is
i a u>e o o/_C*V
where V is an identical vowel.
realistic

Identical to the input or identical to the output? Vowel assimilation will make some sense almost always.* Dissimilation might be a harder sell in this case.

If it's the former, I'd go a bit further even: if you have a change like i…e > e…e, I'd also expect to see some not-quite-all-the-way assimilations like i…a > e…a.

* For some reason this does not work very well for consonants. Assimilation will still always be plausible in clusters, but in the case of consonants separated by a vowel, then dissimilation is maybe more common, both for places and manners of articulation. I guess nasality is the main exception, and then that could be due to transmission by the vowel: a change like "ban > man" is possibly in more detail [ban] > [bãn] > [mãn].

The Caananite languages had vowel dissimilation of the sort a > i /_...a.

[Whimemsz]

That was quick!

Okay, that might actually work. I have also considered debuccalizing /f/ (potential Basque adstrate?) Though despite its rarity, I don't see much issue with /f/ > /p/ if my substrate in question simply pronounces all /f/ as /p/, although for that I'd like a second opinion.

I meant to add that I suspect most /f/ > /h/ is really /ɸ/ > /h/. For sure it was in Japanese and Celtic, and it seems likely also in Polynesian. Spanish has true /f/ > /h/ in initial position, though.

That was also my thinking. I could attribute to a Basque substrate as much as an Innu substrate, much like in Spanish, as Basque also supposedly lacked this phoneme. (And IIRC Basque whalers/traders were active in the region.)

They were, and cool enough, Mi'kmaq actually retains a couple of Basque loanwords (via the Basque-based trade jargon that was used in the region for over a century). Le Jeune reports the Innus using the jargon [including some Innu words] in addressing him when he met with them early on, attesting three Basque-derived words: in asking him "Ania kir Capitana?" "My brother, are you the Captain [= the person in charge here]?" (cf. Basque anaia "brother", kapitaina "captain", Old Innu */kʲi:r/ 2sg) and saying "Ania Capitana ouias amiscou" "My brother, Captain, we will bring you beaver meat" (Old Innu */wi:ja:s/ "flesh" and */amiskʷ/ "beaver") and "Ania Cabana" "My brother, let us go to your cabin" (cf. Basque kabina "cabin of a ship"). By the time of his report a year later (for 1633), thanks to living among the people for all that time and his now significant exposure to their language, he had come to realize that this was a trade jargon and not actual Innu:

I have noticed in the study of their language that there is a certain jargon ["barragoin"] between the French & the Indians, which is neither French nor Indian, & yet when the French use it, they think they are speaking Indian, & the Indians in using it think they are speaking good French. I wrote a few words of it last year that I characterized as Indian words, believing them to be so. For example, the word Ania, which I have mentioned above, is an alien word ["vn mot barbare"], the Indians making use of it on every occasion in speaking to the French, & the French in speaking to the Indians, & all use it to say my brother, but in the real Indian [language] of the Montagnaits, Nichtais [= Old Innu */niʃte:s/] means my eldest brother, Nichun [= Old Innu */niʃi:m/?] my youngest; the word Sagamo is used by only a few here to say Captain; the true word is Oukhimau [= Old Innu */ukʲima:w/ "chief"]; I believe this word, Sagamo, comes from Acadia [cf. Penobscot [Eastern Abenaki] /sɤ̀kəmɤ/ "chief", also the source of English "sagamore"], there are many others like it.

(I will say, though, that f [> ɸ] > h is common enough that you don't really need to fall back on any substrate influence to justify it if you don't want to.)

[Xwtek]

Identical to the input or identical to the output? Vowel assimilation will make some sense almost always.* Dissimilation might be a harder sell in this case.

Identical to input. And my proposal is vowel dissimilation.

If it's the former, I'd go a bit further even: if you have a change like i…e > e…e, I'd also expect to see some not-quite-all-the-way assimilations like i…a > e…a.

I echo Tropylium here, and additionally I would say this seems a little odd a formulation as well. Is this supposed to be with subsequent apocope to obscure the conditioning?

Yes and no. My language originally has penultimate stress (counting by mora, not syllable). Then every vowel that is not stressed is deleted as long as cluster permits. It commonly targets open final vowel. If the final vowel is deleted, the dissimilation creates ablaut. My target language is a not-quite-a-triconsonantal conlang. It's not really a triconsonantal conlang because:

1. Most of the time it affects only a single vowel, as most words are monosyllabic. Only few roots are polysyllabic.
2. There is only a few available ablaut forms per root, relying mostly upon affixes to convey meaning. The ablaut form of ką́s is just ksę́ ksį́ and ksǿ̨.
3. Two words with same consonants may come from a completely different root. For example: ką́s and kóus doesn't come from the same root.
4. This language has extensive vowel and tone inventory. Not to mention the nasal vowels. In total, there is 24 + 112 = 136 nucleus. And vowel carries much more information in this language than most semitic languages.

[Ælfwine]

They were, and cool enough, Mi'kmaq actually retains a couple of Basque loanwords (via the Basque-based trade jargon that was used in the region for over a century). Le Jeune reports the Innus using the jargon [including some Innu words] in addressing him when he met with them early on, attesting three Basque-derived words: in asking him "Ania kir Capitana?" "My brother, are you the Captain [= the person in charge here]?" (cf. Basque anaia "brother", kapitaina "captain", Old Innu */kʲi:r/ 2sg) and saying "Ania Capitana ouias amiscou" "My brother, Captain, we will bring you beaver meat" (Old Innu */wi:ja:s/ "flesh" and */amiskʷ/ "beaver") and "Ania Cabana" "My brother, let us go to your cabin" (cf. Basque kabina "cabin of a ship"). By the time of his report a year later (for 1633), thanks to living among the people for all that time and his now significant exposure to their language, he had come to realize that this was a trade jargon and not actual Innu:

I have noticed in the study of their language that there is a certain jargon ["barragoin"] between the French & the Indians, which is neither French nor Indian, & yet when the French use it, they think they are speaking Indian, & the Indians in using it think they are speaking good French. I wrote a few words of it last year that I characterized as Indian words, believing them to be so. For example, the word Ania, which I have mentioned above, is an alien word ["vn mot barbare"], the Indians making use of it on every occasion in speaking to the French, & the French in speaking to the Indians, & all use it to say my brother, but in the real Indian [language] of the Montagnaits, Nichtais [= Old Innu */niʃte:s/] means my eldest brother, Nichun [= Old Innu */niʃi:m/?] my youngest; the word Sagamo is used by only a few here to say Captain; the true word is Oukhimau [= Old Innu */ukʲima:w/ "chief"]; I believe this word, Sagamo, comes from Acadia [cf. Penobscot [Eastern Abenaki] /sɤ̀kəmɤ/ "chief", also the source of English "sagamore"], there are many others like it.

(I will say, though, that f [> ɸ] > h is common enough that you don't really need to fall back on any substrate influence to justify it if you don't want to.)

Thank you, that is very interesting.

BTW, are there any resources on "Old Innu?" It's difficult enough to find resources on Modern Innu, much less the proto-language, and given the timeframe I am operating with I'd prefer Old Innu. (Though I suppose the difference between Old and Modern Innu isn't great.)

[Whimemsz]

.

[mae]

-

[Ælfwine]

<snip all>

You've exceeded my expectations and then some with all this information you have given me. Curious — are you a scholar in this sort of area? I'll definitely take a look at those resources soon.

The development from Old Norse to Icelandic and Faroese share with Innu also a tendency to palatalize before front vowels, as well as the cluster /sk/ > /ʃ/ in Faroese. I suppose the Faroese path (k > tʃ, sk > ʃ) would be taken here. Innu nouns with -ʷ may get borrowed in as -u (possibly becoming identified with Old Norse ōn-stem feminine nouns with their accusative/dative/genitive sg in -u). Likewise, Old Innu nouns ending in -i might get borrowed as strong masculine ja-stems or weak (j)an-stems due to their nominative.

Though this only works for a language influenced by Innu. If I want a more proper, Michifesque mixed language, is there anything I might want to keep in mind about Innu? Might the phonemes /kʷ/ and /mʷ/ would be added to the phonological inventory, despite being rather phonologically "isolated"? Furthermore, it seems like (modern) Innu lacks the preaspiration of its Cree neighbors, and preaspiration is notably a feature in Old Norse...would it be likely to keep this feature?

I propose something like this, in analogy with how Michif gained its current phonology:

/m n/
/ʰp pʰ p b ʰt tʰ t d ʰk kʰ k g/
/ʰtʃ tʃʰ tʃ dʒ/ (later developments probably)
/f s ʃ h/
/ɾ~r j w/
/l/

/i ĩ u ũ/
/ɪ ɪ̃ ʊ ʊ̃/
/e ẽ ø ø̃ ə ə̃ o õ/
/ɛ ɛ̃ œ œ̃ ɔ ɔ̃/
/a ã/

Words with aspiration (including preaspiration) are preserved from Old Norse. Vowels lose length distinction in favor of a quality distinction, like Icelandic (and as it seems, modern Innu). Nasalization is notably preserved. This still seems like a rather large inventory to me, anything more realistic likely to occur?

[Xwtek]

Weirdly, pharyngeal consonant tend to trigger advancement of the vowel, not retracting it.

[Whimemsz]

Weirdly, pharyngeal consonant tend to trigger advancement of the vowel, not retracting it.

What are some examples? This is certainly not the case in, say, Arabic or Chilcotin or Spokane, where pharyngeals/pharyngealized consonants result in vowel lowering, backing, and/or -ATR (or similar).

Suppose a language allows a pharyngeal approximant in both onset and coda position, and you want to get rid of it by turning it into something else. Vʕ > Vˤ is very trivial, but what happens with pharyngeals in the onset? Abkhaz apparently had a change ʕ > j, but what else can happen? If you have something like ʕ > ʔ can the pharyngealization transfer to the following vowel? I'm not too clear on what the diachronics of these sounds are really.

Pharyngealization can definitely transfer to the vowel, very easily. Pharyngealization is also very commonly subject to spreading over multiple syllables (like nasalization) or to metathesis (via pharyngeal spread and then loss or modification of the original segment). Some Salishan languages, partly because of their crazy consonant clusters, have pharyngeal metathesis (from an unstressed root to a stressed suffix) to a particularly fun degree, e.g., Colville has /pʕat/ "boil" --> /s-ən-pə-pt-ʕálaʔqʷ/ "dumpling" (somewhat reasonable) but also /q’ʷʕáj/ "black" --> /q’əj-lstʃʕát/ "his clothes are dirty" and /pʕáw/ "he ran down" --> /pw-ən-tsʕát-əlx/ "they make noise running down."

<snip all>

You've exceeded my expectations and then some with all this information you have given me. Curious — are you a scholar in this sort of area? I'll definitely take a look at those resources soon.

The development from Old Norse to Icelandic and Faroese share with Innu also a tendency to palatalize before front vowels, as well as the cluster /sk/ > /ʃ/ in Faroese. I suppose the Faroese path (k > tʃ, sk > ʃ) would be taken here. Innu nouns with -ʷ may get borrowed in as -u (possibly becoming identified with Old Norse ōn-stem feminine nouns with their accusative/dative/genitive sg in -u). Likewise, Old Innu nouns ending in -i might get borrowed as strong masculine ja-stems or weak (j)an-stems due to their nominative.

Though this only works for a language influenced by Innu. If I want a more proper, Michifesque mixed language, is there anything I might want to keep in mind about Innu? Might the phonemes /kʷ/ and /mʷ/ would be added to the phonological inventory, despite being rather phonologically "isolated"? Furthermore, it seems like (modern) Innu lacks the preaspiration of its Cree neighbors, and preaspiration is notably a feature in Old Norse...would it be likely to keep this feature?

I propose something like this, in analogy with how Michif gained its current phonology:

/m n/
/ʰp pʰ p b ʰt tʰ t d ʰk kʰ k g/
/ʰtʃ tʃʰ tʃ dʒ/ (later developments probably)
/f s ʃ h/
/ɾ~r j w/
/l/

/i ĩ u ũ/
/ɪ ɪ̃ ʊ ʊ̃/
/e ẽ ø ø̃ ə ə̃ o õ/
/ɛ ɛ̃ œ œ̃ ɔ ɔ̃/
/a ã/

Words with aspiration (including preaspiration) are preserved from Old Norse. Vowels lose length distinction in favor of a quality distinction, like Icelandic (and as it seems, modern Innu). Nasalization is notably preserved. This still seems like a rather large inventory to me, anything more realistic likely to occur?

Not a scholar, no, just an interested random person like most people here. I dunno, that seems reasonably realistic to me for a Michif-y outcome? The mixed Michif phonology is pretty large, although some phonemic distinctions found in most dialects of French, including other dialects of Canadian French, seem to have been lost in Métis French, so that kept it from getting too out of hand. I imagine preaspiration probably would remain, although the pressure from neighboring varieties of CMN which also lost it might eventually have an effect (or alternately, Norse-Innu could have an effect on them).

At the time they'd be meeting Old Norse speakers, Innu speakers would still have final vowels after /-kw/ and /-mw/, so depending on whether you have sound changes happen like they did in our timeline you might not need to worry about /kʷ/ and /mʷ/. Also, since you mention noun endingns, all Old Innu singular nouns would have ended in either */-a/ (animate; plural */-aki/, obviative */-ahi/) or */-i/ (inanimate; plural */-ahi/)

[Nortaneous]

Weirdly, pharyngeal consonant tend to trigger advancement of the vowel, not retracting it.

What are some examples? This is certainly not the case in, say, Arabic or Chilcotin or Spokane, where pharyngeals/pharyngealized consonants result in vowel lowering, backing, and/or -ATR (or similar).

This is sometimes cited as the difference between pharyngeals and epiglottals: pharyngeals pull vowels toward /ɑ/, whereas epiglottals pull them toward /æ/.

[Xwtek]

Weirdly, pharyngeal consonant tend to trigger advancement of the vowel, not retracting it.

What are some examples? This is certainly not the case in, say, Arabic or Chilcotin or Spokane, where pharyngeals/pharyngealized consonants result in vowel lowering, backing, and/or -ATR (or similar).

Actually, Arabic IS the example of this phenomenon (or at least Cairene Arabic). https://escholarship.org/content/qt50v3 ... f?t=nmp4q5. Note that pharyngealized consonants DO retract the vowel, but the pharyngeal consonants itself don't. For the more extreme example, Avar rounded back vowel is pronounced as rounded front vowel next to pharyngeal consonant: /goħ/ > [gøħ].

[Xwtek]

Is it realistic to have /q/ > /k/ if you have a pharyngeal consonant?

[bradrn]

Is it realistic to have /q/ > /k/ if you have a pharyngeal consonant?

According to the Index Diachronica, it’s happened in Cypriot Arabic. In general though, I don’t see anything which would make this particularly unrealistic.

(BTW, are you the same person formerly known as Akangka?)