Zompist Bboard Again

How realistic is it for the high vowels /i y u/ to neutralize to [ɨ] in unstressed syllables, without the mid vowels /e ø ɔ/ doing something similar and merging to [ə]? I already have /a/ > [ə] in unstressed syllables instead.

On a related note, I also have processes of vowel reduction and deletion which occur only in the syllable following the stressed syllable, which I think I recall happening somewhere in Germanic, but I'm not totally sure?

The mid vowels could merge into a slightly higher vowel than /a/

Max1461 wrote: ↑Thu Aug 09, 2018 11:06 am On a related note, I also have processes of vowel reduction and deletion which occur only in the syllable following the stressed syllable, which I think I recall happening somewhere in Germanic, but I'm not totally sure?

I don't know about natlangs but that sounds perfectly reasonable. I've always called this a "Debra" shift, following the pattern where the name Deborah appears as Debra , but never *Debor.

I'm pretty sure it occurred many times in Indo-European

Hello again!

An incatena forumite returning with the same question and a few others:

Firstly, how realistic would it be for a language with a very robust voicing distinction and a secondary articulation series to lose both of these completely and quickly? I could imagine that secondary articulation - uvularisation in my case - could become a difference in quality of surrounding vowels, doubling the vowel inventory from 4 to 8, but what about voicing?

Secondly, and in relation to the above, how feasible would it be to shift /ɣ~ʁ/ into /j/ before front high vowels and /ŋ/ everywhere else?

Finally, how plausible does it sound to have some pervasive and strong diachronic post-nasal fortition and voicing where every nasal + consonant cluster shifts to one of /mb nd nʲdʒ ŋg/ and eventually these clusters are reanalysed as prenasalised consonants, giving rise to a new series? What could trigger such a reanalysis? Also, as a consequence, the language would have severely limited nasal+consonant clusters, with most nasal+voiced plosive at the same POA sequences being found in recent loanwords and only a few in archaic ossified grammatical constructs - does that ever happen with natlangs?

Knit Tie wrote: ↑Fri Aug 10, 2018 1:22 pm

Finally, how plausible does it sound to have some pervasive and strong diachronic post-nasal fortition and voicing where every nasal + consonant cluster shifts to one of /mb nd nʲdʒ ŋg/ and eventually these clusters are reanalysed as prenasalised consonants, giving rise to a new series?

I've used something like this for all but nasal-nasal, and I'm sure that could also happen ..... But I haven't looked for natlang parallels. English has ml>mbl,additionally also nl>ndl,I think. Latin mr>mbr.with thesey ou could then lose the final approximant .
.

What could trigger such a reanalysis?

as long as they either:

1)contrast with nasal+stop sequences, or
2) are the only otherwise closed syllables in the language,
It would make sense to analyse mb, nd, etc as singletons.

Also, as a consequence, the language would have severely limited nasal+consonant clusters, with most nasal+voiced plosive at the same POA sequences being found in recent loanwords and only a few in archaic ossified grammatical constructs - does that ever happen with natlangs?

I don't follow here. Please explain.

Pabappa wrote: ↑Fri Aug 10, 2018 2:41 pm

Also, as a consequence, the language would have severely limited nasal+consonant clusters, with most nasal+voiced plosive at the same POA sequences being found in recent loanwords and only a few in archaic ossified grammatical constructs - does that ever happen with natlangs?

I don't follow here. Please explain.

Basically, are there any coda- and cluster-permitting natlangs that limit what can cluster with a nasal? So any languages that allow things like VC1C2V, where either C1 can't be a nasal or C2 can be only something particular if C1 is a nasal?

Also, if a language disallows any intital consonant clusters but has /mb/ and the like word-initially, would it be sufficient reason to postulate the existence of a prenasalised series? What about if voiced stops are only ever found post-nasally in the language?

Knit Tie wrote: ↑Fri Aug 10, 2018 1:22 pmFirstly, how realistic would it be for a language with a very robust voicing distinction and a secondary articulation series to lose both of these completely and quickly? I could imagine that secondary articulation - uvularisation in my case - could become a difference in quality of surrounding vowels, doubling the vowel inventory from 4 to 8, but what about voicing?

It's very possible. The most obvious thing to do with voicing would be tonogenesis. Another option would be P > Pʰ, B > P (or a variant: P > Pʰ, B > Pʼ or ɓ). A third option, of course, would simply be P B > P.

Secondly, and in relation to the above, how feasible would it be to shift /ɣ~ʁ/ into /j/ before front high vowels and /ŋ/ everywhere else?

The first is very feasible. The second...well, I've seen it happen in the other direction, for what that's worth.

Finally, how plausible does it sound to have some pervasive and strong diachronic post-nasal fortition and voicing where every nasal + consonant cluster shifts to one of /mb nd nʲdʒ ŋg/ and eventually these clusters are reanalysed as prenasalised consonants, giving rise to a new series? What could trigger such a reanalysis?

Perfectly plausible. If I'm not much mistaken, that's how analysis of prenasalized consonants usually works, since acoustically they're quite similar to nasal-plosive clusters.

Also, as a consequence, the language would have severely limited nasal+consonant clusters, with most nasal+voiced plosive at the same POA sequences being found in recent loanwords and only a few in archaic ossified grammatical constructs - does that ever happen with natlangs?

I'm not very familiar with languages with prenasalized consonants, but broadly speaking phonemes or phoneme sequences found only in select circumstances, especially loanwords, are definitely a thing.

can pharyngealized/emphatic consonants turn into ejectives?

Knit Tie wrote: ↑Fri Aug 10, 2018 2:54 pm

Pabappa wrote: ↑Fri Aug 10, 2018 2:41 pm

Also, as a consequence, the language would have severely limited nasal+consonant clusters, with most nasal+voiced plosive at the same POA sequences being found in recent loanwords and only a few in archaic ossified grammatical constructs - does that ever happen with natlangs?

I don't follow here. Please explain.

Basically, are there any coda- and cluster-permitting natlangs that limit what can cluster with a nasal? So any languages that allow things like VC1C2V, where either C1 can't be a nasal or C2 can be only something particular if C1 is a nasal?

Well the part I didnt understand was how you'd get those sequences in loanwords. If your language has a series of prenasalized stops /mb nd ŋg/ etc, I'd expect that foreign loanwords with sequences would be borrowed in using the prenasals, so there would be no contrast. Some languages do indeed distinguish prenasal /mb nd ŋg/ from two-consonant sequences, but I wouldnt think a contrast would arise just from foreign loans, unless this language is mora-timed and borrows syllable length along with the phonemes.

Also, if a language disallows any intital consonant clusters but has /mb/ and the like word-initially, would it be sufficient reason to postulate the existence of a prenasalised series? What about if voiced stops are only ever found post-nasally in the language?

Yes, definitely (to both). But Im still curious if youre planning on having a contrast between prenasalized stops and nasal+stop sequences.

÷÷÷÷÷÷÷÷÷÷÷
some snd changes Ive done .... note though that I dont generally look for natlang precedence, so these are all changes that just feel right to me, and may not have actually occcurred anywhere. Also, these changes do not all occur in the same language, hence why they seem to contradict each other:

/sm/ > /hm/ > /mh/ > /mf/ > /mp/ ... and likewise for other similar sequences ... this does not shift to /mb/, but to do so would be trivial
/ml mr/ > /mbl mbr/> /mb mb/
/mm nn ŋŋ/ > /mb nd ŋg/
/mʲ nʲ ŋʲ/ > /mb ndʒ ndʒ/
/mʷ nʷ ŋʷ/ > /mbʷ ŋgʷ ŋgʷ/
/nʷ/ > /ndʷ/ > /nd/
/mʕ / > /mb /

bbbosborne wrote: ↑Fri Aug 10, 2018 8:35 pm can pharyngealized/emphatic consonants turn into ejectives?

Well, the Semitic emphatics were almost certainly originally ejective before becoming pharyngealized in Aramaic and Arabic, and ejective to pharyngealized is attested in Caucasian as well. Can it go the other way? I hope so, because I used that change myself recently.

I've also used this shift. As above, I go with what feels right rather than looking for natlangs that've done it, but I always think about probabilities. For this shift, it m8ght work best if you have a contrast of voiceless vs aspirated stops, where the plain voiceless series takes on pharyngealization in some contexts, and a later shift creates ejectives. Since asps cannot turn into ejectives, they will be immune to the shift.. then any remaining phars can either become plain, or shift to ejectives by analogy.

Pabappa wrote: ↑Fri Aug 10, 2018 10:59 pm I've also used this shift. As above, I go with what feels right rather than looking for natlangs that've done it, but I always think about probabilities. For this shift, it m8ght work best if you have a contrast of voiceless vs aspirated stops, where the plain voiceless series takes on pharyngealization in some contexts, and a later shift creates ejectives. Since asps cannot turn into ejectives, they will be immune to the shift.. then any remaining phars can either become plain, or shift to ejectives by analogy.

Which reminds me that something that is attested is plain voiceless > ejective, so yes, if your non-pharyngealized plain voiceless plosives develop aspiration to heighten contrast with the pharyngealized plosives, then your pharyngealized plosives lose their pharyngealization, then become ejective to heighten contrast with the aspirated unvoiced plosives, that would be a very valid method of pharyngealized > ejective. Not sure it will work for fricatives, though, since aspirated fricatives are pretty rare (but then again, so are ejective fricatives, found chiefly in the Caucasus and PNW).

Given bog standard /a o u e i/ and /w/ in the onset, is it plausible that
w > ∅ / _o
w > v / _{a u e i}
with the latter perhaps as two seperate sound changes - / _u and / _{a e i} ? Order doesn't matter.

Pabappa wrote: ↑Fri Aug 10, 2018 8:56 pm Well the part I didnt understand was how you'd get those sequences in loanwords. If your language has a series of prenasalized stops /mb nd ŋg/ etc, I'd expect that foreign loanwords with sequences would be borrowed in using the prenasals, so there would be no contrast. Some languages do indeed distinguish prenasal /mb nd ŋg/ from two-consonant sequences, but I wouldnt think a contrast would arise just from foreign loans, unless this language is mora-timed and borrows syllable length along with the phonemes.

Ah, I see. I'm not talking about the language having nasal+consonant sequences in loanwords, I'm talking about it not allowing things like /mr/ or /nl/ or even /ns/ and /nt/ at all in native vocabulary, in fact, no consonant clusters at all are possible with nasals in initial position, as historically, all such clusters became /mb/ etc. and later on these sequences got reanalysed as single consonants. So when you have something like /nt/, you know it's a loanword. This brutal degree of post-nasal fortition is supposed to be rather unique and highly indicative of the language, and I'm wondering if something like this has ever been attested in real life.

Pabappa wrote: ↑Fri Aug 10, 2018 8:56 pm But Im still curious if youre planning on having a contrast between prenasalized stops and nasal+stop sequences.

Definitely no, as I've said above - I'm actually planning on having no nasal+stop sequences at all.

Zju wrote: ↑Sat Aug 11, 2018 11:48 am Given bog standard /a o u e i/ and /w/ in the onset, is it plausible that
w > ∅ / _o
w > v / _{a u e i}
with the latter perhaps as two seperate sound changes - / _u and / _{a e i} ? Order doesn't matter.

I would expect w > ∅ /_u as well, if you have w > ∅ / _o, but I could maybe see w > v /_u by dissimilation

Why a front vowel in the first place though?

mèþru wrote: ↑Sat Aug 11, 2018 1:19 pm Why a front vowel in the first place though?

∅, the empty set symbol, easily confused with [ø], the mid front rounded vowel.

Ah, they look more different in the font I usually see them in

Zju wrote: ↑Sat Aug 11, 2018 11:48 am Given bog standard /a o u e i/ and /w/ in the onset, is it plausible that
w > ∅ / _o
w > v / _{a u e i}
with the latter perhaps as two seperate sound changes - / _u and / _{a e i} ? Order doesn't matter.

w//_[ou] happened in a lot of Indo-European varieties. If w/v/_u happened before w//_o, I'd certainly buy it as dissimilation.