Sound Change Quickie Thread

[bbbosborne]

not sure if this belongs in a separative thread, but:
how do stress rules change over time?

[Travis B.]

not sure if this belongs in a separative thread, but:
how do stress rules change over time?

Take for instance a language with antepenultimate stress. Then consider that some final vowels are lost. As a result one gets a language where stress can be either antepenultimate or penultimate.

Another case to consider is a language with root-initial stress. Then consider the case where words are borrowed from another language with a different stress paradigm. Then the stress rules can follow those of native words or those of loanwords.

Yet another case to consider is a language with a syllable weight-based stress paradigm. Then consider the case that the syllable weight system collapses, e.g. through the loss of phonemic vowel length. As a result stress becomes phonemic, because it is no longer predictable.

[Linguoboy]

not sure if this belongs in a separative thread, but:
how do stress rules change over time?

And then of course there are substratal and superstratal effects, e.g. French stress shifting to word initial under the influence of Frankish (and then subsequently shifting to the end of each breath group).

Heavy syllables can attract stress. Stress is fixed on the initial syllable in modern Goidelic outside of Munster Irish where a long vowel in the second syllable causes a stress shift. E.g. Munster fadó /fəˈdoː/ vs Connacht /ˈfadoː/, Ulster /ˈfado/. Obviously, any change which alters syllable weight can also lead to stress changes, e.g. gadaighe /ˈgadaɣʹi/ "thief" > gadaí /ˈgadiː/ > [Munster] /gəˈdiː/.

[Tropylium]

Non-phonemic longer vowels, such as more open vowels (/a/ [aˑ] versus /i/ [i]) or vowels in open syllables can also be stress attractors (at least as long as there are no phonemic long vowels).

[Knit Tie]

What are some of the ways a language can develop an initial-only velar nasal? Are there any phonemes that love turning into one?

[statelessnation]

redacted

[Knit Tie]

Does rhinoglottophilia make /ʕ/ to eng a viable sound shift, actually?

[mèþru]

Yes, but the specific scenario by statelessnation seams pretty unlikely to me yet still plausible.

[Pabappa]

Does rhinoglottophilia make /ʕ/ to eng a viable sound shift, actually?

I wrote once that it happened in Hebrew, giving the diminutive "Yankl" from Ya‘aqōv , but it seems this was a case of sound substitution rather than a true sound change.

[Travis B.]

One route I thought of is to have initial-only consonant cluster reduction, and have the initial clusters /kn/ and /gn/ reduce to [ŋ], resulting in an initial-only /ŋ/ phoneme.

[missals]

What are some of the ways a language can develop an initial-only velar nasal? Are there any phonemes that love turning into one?

This can actually just happen by itself. Some languages, I think in particular some of the Samoyedic languages, have undergone a sound change where all vowel-initial words gained an initial [ŋ]. In some this was more conditioned, with initial front vowels developing an initial [ɲ].

I wrote once that it happened in Hebrew, giving the diminutive "Yankl" from Ya‘aqōv , but it seems this was a case of sound substitution rather than a true sound change.

I've heard people say it was just a "sound substitution" or "pronunciation change", but I don't really know what that means, and I don't think it's meaningfully different from a sound change in this context. [ʕ] > [ŋ] is itself a plausible sound change; in addition to considering rhinoglottophilia as a theoretical possibility, I recall that Cyrus H. Gordon, scholar of Middle Eastern languages and one of the decipherers of Ugaritic, noted that learners of Arabic often substitute [ŋ] for [ʕ].

[dɮ the phoneme]

This can actually just happen by itself. Some languages, I think in particular some of the Samoyedic languages, have undergone a sound change where all vowel-initial words gained an initial [ŋ].

If you're referring to Tundra Nenets, I believe some dialects lost initial velars nasals, creating the only vowel-initial words, rather than vice-versa.

[gach]

This can actually just happen by itself. Some languages, I think in particular some of the Samoyedic languages, have undergone a sound change where all vowel-initial words gained an initial [ŋ].

If you're referring to Tundra Nenets, I believe some dialects lost initial velars nasals, creating the only vowel-initial words, rather than vice-versa.

Tagging nasals in front of vowel initial words is certainly a northern Samoyedic thing. I'm travelling now, though, so someone else will have to check the exact distribution for you.

[Tropylium]

It goes in a couple of phases. The following outline is adapted from a 1975 paper by Sammallahti:

1) some distinct glide losses and epentheses in the early Northern Samoyedic dialects (e.g. Nenets-Enets *jilɛ- 'to live' versus Nganasan *elɛ-; Nen-En *ïnta 'bow' versus Nganasan *jïntå)
2) general epenthesis of *ŋ before all vowel-initial words (some people have thought this was actually a retention from Proto-Uralic, but probably not, since it affects also words with secondary vocalic onset due to the previous)
3) palatalization of *ŋ to *ŋʲ before front vowels;
4a) in Tundra Enets, loss of all *ŋ, including *ŋʲ; initial ŋ- then reintroduced by loans from Tundra Nenets
4b) in Forest Enets, loss of all *ŋ except *ŋʲa > *nʲɑ > /nɑ/ (versus ŋʲa > *ŋʲæ > *æ > /ɛ/ in Tundra Enets)
4c) in Nenets and Nganasan, merger of *ŋʲ into /nʲ/ always
5) secondary loss of retained /ŋ/ in far western dialects of Tundra Nenets.

[dɮ the phoneme]

So, I have a proto-lang with a very constrained root structure of CV(C)CV, and I'm trying to develop contrastive palatalization. Presently, I have palatalization of consonants before front vowels, followed by elision of vowels in the syllable following stress —which turns out to be morphologically predictable, and is usually the last syllable of the root. This results in a CV(C)C root structure, with the occasional morphologically predictable final vowel, and contrastive palatalization in the final consonants. The problem is, palatalization is still only marginally contrastive in the initial consonant, owing to a couple of vowel shifts, which seems like an odd situation to me. Aside from some widespread front-back merger of vowels, which I don't really want to do, I'm not quite sure how to make palatalization contrastive root-initially. Any ideas?

[gach]

Aside from some widespread front-back merger of vowels, which I don't really want to do, I'm not quite sure how to make palatalization contrastive root-initially. Any ideas?

The simplest way would be to introduce new roots to the language that don't respect the inherited palatal distribution. Either have a new layer of loans or invent some onomatopoeia. You could also investigate a more limited vowel merger that would boost the palatal contrast but still leave the vowel system largely unaffected.

[sangi39]

So, I have a proto-lang with a very constrained root structure of CV(C)CV, and I'm trying to develop contrastive palatalization. Presently, I have palatalization of consonants before front vowels, followed by elision of vowels in the syllable following stress —which turns out to be morphologically predictable, and is usually the last syllable of the root. This results in a CV(C)C root structure, with the occasional morphologically predictable final vowel, and contrastive palatalization in the final consonants. The problem is, palatalization is still only marginally contrastive in the initial consonant, owing to a couple of vowel shifts, which seems like an odd situation to me. Aside from some widespread front-back merger of vowels, which I don't really want to do, I'm not quite sure how to make palatalization contrastive root-initially. Any ideas?

There's also the option of "umlaut" followed by vowel mergers. So you get something like this:

Code: Select all

Initial root types (25) + Changes

CiCi CeCi CuCi CoCi CaCi > CʲiCʲi CʲeCʲi CuCʲi CoCʲi CaCʲi > CʲiCʲi CʲiCʲi CyCʲi CøCʲi CəCʲi > CʲiCʲə CʲiCʲə CiCʲə CeCʲə CəCʲə > CʲiCʲ CʲiCʲ CiCʲ CeCʲ CəCʲ
CiCu CeCu CuCu CoCu CaCu > CʲiCu  CʲeCu  CuCu  CoCu  CaCu  > CʲʉCu  CʲɵCu  CuCu  CuCu  CəCu  > CʲuCə  CʲoCə  CuCə  CuCə  CəCə  > CʲuC  CʲoC  CuC  CuC  CəC
CiCe CeCe CuCe CoCe CaCe > CʲiCʲe CʲeCʲe CuCʲe CoCʲe CaCʲe > CʲiCʲe CʲeCʲe CuCʲe CoCʲe CaCʲe > CʲiCʲə CʲeCʲə CuCʲə CoCʲə CaCʲə > CʲiCʲ CʲeCʲ CuCʲ CoCʲ CaCʲ
CiCo CeCo CuCo CoCo CaCo > CʲiCo  CʲeCo  CuCo  CoCo  CaCo  > CʲiCo  CʲeCo  CuCo  CoCo  CaCo  > CʲiCə  CʲeCə  CuCə  CoCə  CaCə  > CʲiC  CʲeC  CuC  CoC  CaC
CiCa CeCa CuCa CoCa CaCa > CʲiCa  CʲeCa  CuCa  CoCa  CaCa  > CʲeCa  CʲɛCa  CoCa  CɔCa  CaCa  > CʲeCə  CʲaCə  CoCə  CaCə  CaCə  > CʲeC  CʲaC  CoC  CaC  CaC

Resulting root types (17):

CʲiCʲ CʲeCʲ

CʲiC CʲeC CʲuC CʲoC CʲaC

CiCʲ CeCʲ CuCʲ CoCʲ CaCʲ CəCʲ

CuC  CuC  CaC CəC

So you have four broad classes of roots, a "heavily front" one (both consonants palatalised with a front vowel), a "heavily back" one (no palatalisation at all, with a non-front vowel), and then two other classed based on whether the first or second consonant is palatalised.

There's probably something that could be done with the introduction of long vowels which you could do to keep the number of root types from decreasing so much, with long vowels having different outcomes from their short counterparts (lengthen the first vowel in CVCV roots, have vowels change as needed, then have length become contrastive by deleting syllable-final nasals, maybe?), but the above is the first thing that came to mind.

[mèþru]

You could also do some kind of consonant-umlaut or consonant harmony
You could also have palatalisation triggered by preceding vowels, especially by diphthongisation (and the diphthongs could then simplify).

[dɮ the phoneme]

The simplest way would be to introduce new roots to the language that don't respect the inherited palatal distribution. Either have a new layer of loans or invent some onomatopoeia.

I think this might be the simplest approach. Even if I find a way to internally derive them, I'll probably end up doing this too.

There's also the option of "umlaut" followed by vowel mergers.

I've already had umlaut occur, resulting in some front-rounded vowels in roots where the final consonant is palatalized. The current vowel system is something like this:

Code: Select all

i		y		ɨ	ɨ̃	u
		ø	ø̃				õ
æ	æ̃					ɔ	ɔ̃
				a	ã

Which is pretty unstable looking. I was hoping to simplify it in different ways in the daughter langs, but I don't see any obvious mergers that would expand the distribution of initial palatals.

Unrelated addendum:

Is something like [ɨ] → ∅ /[+sonorant]_, except _[+sonorant] realistic, given that [ɨ] only occurs in unstressed syllables?

[mèþru]

What about infixation or duplifxiation before the sound changes? Then you could reduce to a single syllable retaining some properties of both syllables.