Sound Change Quickie Thread

[dɮ the phoneme]

I think nasals could devoice allophonically and thus permit the shift.

Would this really be a necessary precondition? Since nasals are voiced, they can (phonetically) show tone distinctions themselves, so on second thought it might not matter whether the tone-conditionaing elements directly abut the vowel or a nasal/approximate. The only other kind of coda clusters permitted in this language are -PC and -sC; it seems to me that maybe glottalization could spread backwards through those more easily than voice?

[Pabappa]

I think nasals could devoice allophonically and thus permit the shift.

Would this really be a necessary precondition? Since nasals are voiced, they can (phonetically) show tone distinctions themselves, so on second thought it might not matter whether the tone-conditionaing elements directly abut the vowel or a nasal/approximate. The only other kind of coda clusters permitted in this language are -PC and -sC; it seems to me that maybe glottalization could spread backwards through those more easily than voice?

I like to work in baby steps. Specifying voiceless allophones for sonorants helped me think of a very important sound change in Poswa, so that's how I always do things now. But you're right that nasals can carry tone .... Lithuanian has a pitch accent system in which the sonorants in sequences like /im/ can carry distinctive tones; that's where the Lithuanian m̃ comes from. I'm still a bit cautious though about a language with tone contrasts in syllables that end in a cluster. I am probably going to do this in a conlang myself, but the "tones" will also be distinguished by other things such as length, color, and possibly pharyngealization. Even so, I think there are at least some languages that permit tone + coda cluster, so it's not entirely unattested. (I dont know if Lithuanian does .... its tone contrasts are basically only for long vowels, and I dont know if you can have a long vowel plus a cluster in a word-final syllable.)

[dɮ the phoneme]

I'm still a bit cautious though about a language with tone contrasts in syllables that end in a cluster.

At least Haida does it IIRC, as well as Gaahmg appearently.

[Knit Tie]

Can I say thay r-coloured vowels, after /ɻ/ becomes /j/, become Vɪ̯ everywhere except word-finally, where they become jV instead? Maybe just break the r-coloured vowels into ɻV sequences word-finally beforehand?

Actually, I might just delete /ɻ/ in all codas and turn all rhotic vowels into regular ones. The idea I had, with jV and Vj sequences, chokes the lexicon in yods and I don't like how it sounds.

[dɮ the phoneme]

Can I say thay r-coloured vowels, after /ɻ/ becomes /j/, become Vɪ̯ everywhere except word-finally, where they become jV instead? Maybe just break the r-coloured vowels into ɻV sequences word-finally beforehand?

I would believe this, at least by way of ɻ > ɹ > j. And I have [fɹʷənətʃʷɚ] for 'furniture', so I'd also buy V˞ > ɻV.

[Pabappa]

I'm surprised this isn't more common in conlangs - there are dialects of English with this contrast, and in some without it, the interdentals often undergo phonetic fortition to stops that contrast with the alveolar ones.

I have dentals in a few languages.... always derived from earlier /t/ vs /t_j/.... But I mostly show the languages with simpler phonologies.

[Nortaneous]

Maybe just break the r-coloured vowels into ɻV sequences word-finally beforehand?

there's a dialect of Mandarin where erhua becomes a medial (infixed) retroflex lateral

[Knit Tie]

Okay, after some deliberation, decided to go with my original idea about rhotacised vowels turning into Vj and jV sequences and creating a yod deluge, which leads to palatalisation. On top of that, I've decided to stick on some voiced affricate lenition, so tell me how do these sound changes look like:

t̪j → t̪s̪
d̪j → d̪z̪ → ̪z
tj → tʃ
dj → dʒ → j

Plus, /xj/ becomes a /ʃ/.

[StrangerCoug]

Those all look plausible to me.

[dɮ the phoneme]

Could voiced obstruents in the coda take on breathy voice, then spread it to a preceding vowel?

[Zju]

Could voiced obstruents in the coda take on breathy voice

Not at all sure about that, especially unconditionally. Maybe they could turn to /ɦ/ in some environments?

[coda position] breathy voice, then spread it to a preceding vowel?

As about natural as it gets, IMO.

[akam chinjir]

Suppose that a language allows all the following nasal+plosive clusers between vowels:

• mp
• np
• mt
• nt
• mk
• ŋk

What I'm wondering is whether it would make sense for homorganic nasals to be preserved, but nonhomorganic nasals to drop, nasalising the preceding vowel:

• ampa → ampa
• anpa → ãpa
• amta → ãta
• anta → anta
• amka → ãka
• aŋka → aŋka

[Xwtek]

Could voiced obstruents in the coda take on breathy voice, then spread it to a preceding vowel?

No, except you allow voiced obstruents in the onset to also take on breathy voice. Then the situation is like Javanese.

[StrangerCoug]

Suppose that a language allows all the following nasal+plosive clusers between vowels:

• mp
• np
• mt
• nt
• mk
• ŋk

What I'm wondering is whether it would make sense for homorganic nasals to be preserved, but nonhomorganic nasals to drop, nasalising the preceding vowel:

• ampa → ampa
• anpa → ãpa
• amta → ãta
• anta → anta
• amka → ãka
• aŋka → aŋka

I buy it. There's a part of me that expects allophonic nasal vowels before the remaining nasal consonants, but if you want them to remain oral, you can see what ideas you can get from French.

[missals]

Suppose that a language allows all the following nasal+plosive clusers between vowels:

• mp
• np
• mt
• nt
• mk
• ŋk

What I'm wondering is whether it would make sense for homorganic nasals to be preserved, but nonhomorganic nasals to drop, nasalising the preceding vowel:

• ampa → ampa
• anpa → ãpa
• amta → ãta
• anta → anta
• amka → ãka
• aŋka → aŋka

I would expect exactly the reverse, actually - homorganic nasals are less phonologically "marked"; i.e. they're not encoding a POA contrast and they're more "reduced", having gone through more assimilatory processes than the nonhomorganic nasals. The nonhomorganic nasals would almost certainly be diachronically newer in origin and would not have undergone assimilation yet. The next logical step for the older, homogranic set of coda nasals would be to project nasality onto the preceding vowel and disappear as separate units. The next logical step for the newer, nonhomorganic set of nasals would be to become homorganic with the following consonant.

As a set of diachronic sound changes, it would look roughly like this (unsure about the exact notation):

1. VN^[POAσ]C^[POAσ] > ṼC
2. N^[POA1]C^[POA2] > N^[POA2]C^[POA2]

That is:

1. Any VNC sequence where the N and C share the same place of articulation become ṼC
2. All nasals assimilate in place to the following consonant

[akam chinjir]

Thanks StrangerCoug and missals!

And missals, that's a great idea.

[bradrn]

A few questions:

1. Can the glottal stop do anything other than disappearing or turning into /h/?
2. Is it plausible to have x, ɣ → j / unconditionally? If not, what conditions could this occur in (if any)?
3. In a syllable-timed language without any sort of contrastive stress, it plausible to have [+short] C C → [+long] C / when Cs are the same? e.g atta → aːta, mənna → məːna
4. In the same language, is it plausible to have [+short] {w,y,r} [+short] → [+long] / when vowels are the same? e.g. tæjæs → tæːs, ʔewe → ʔeː

[Whimemsz]

Can the glottal stop do anything other than disappearing or turning into /h/?

Not much, really...glottal stops are pretty much the end of the line. Maybe turning into epiglottals or something very close, which could then maybe do other things? But I'm not aware offhand of any actual examples of that.

Is it plausible to have x, ɣ → j / unconditionally? If not, what conditions could this occur in (if any)?

Probably implausible in one step, although you could do it with multiple steps (e.g., with [h] as an intermediary). There are conditions where it would be plausible in one step, though, most obviously when in contact with a front vowel. Some Romance languages also probably had a change of coda [x] > [j] (it's the most reasonable interpretation of what happened with, say, noctem >> *[noxte] > *[nojte] >> Fr. nuit, Pt. noite, Sp. noche).

In a syllable-timed language without any sort of contrastive stress, it plausible to have [+short] C C → [+long] C / when Cs are the same? e.g atta → aːta, mənna → məːna

In the same language, is it plausible to have [+short] {w,y,r} [+short] → [+long] / when vowels are the same? e.g. tæjæs → tæːs, ʔewe → ʔeː

Yes to both of these

[akam chinjir]

Can the glottal stop do anything other than disappearing or turning into /h/?

Not much, really...glottal stops are pretty much the end of the line. Maybe turning into epiglottals or something very close, which could then maybe do other things? But I'm not aware offhand of any actual examples of that.

Glottal stops can leave tone behind; glottalisation as well?

[bradrn]

Can the glottal stop do anything other than disappearing or turning into /h/?

Not much, really...glottal stops are pretty much the end of the line. Maybe turning into epiglottals or something very close, which could then maybe do other things? But I'm not aware offhand of any actual examples of that.

Pity… I was hoping to do something a bit more interesting. As akam chinjir mentions, you can do a sound change next to a glottal stop and then delete the glottal stop, but that isn’t quite the same thing.

Is it plausible to have x, ɣ → j / unconditionally? If not, what conditions could this occur in (if any)?

Probably implausible in one step, although you could do it with multiple steps (e.g., with [h] as an intermediary). There are conditions where it would be plausible in one step, though, most obviously when in contact with a front vowel. Some Romance languages also probably had a change of coda [x] > [j] (it's the most reasonable interpretation of what happened with, say, noctem >> *[noxte] > *[nojte] >> Fr. nuit, Pt. noite, Sp. noche).

Why do you think this is implausible? Is there any particular phonetic reason why? I would have thought that it’s a simple lenition from fricative to approximant.

In a syllable-timed language without any sort of contrastive stress, it plausible to have [+short] C C → [+long] C / when Cs are the same? e.g atta → aːta, mənna → məːna

In the same language, is it plausible to have [+short] {w,y,r} [+short] → [+long] / when vowels are the same? e.g. tæjæs → tæːs, ʔewe → ʔeː

Yes to both of these

Thanks!