#phonological evolution

[paizau]

Current palatalization situation of Kshafa stops:

C+j clusters evolved into palatalized consonants twice.

In the first palatalization the velar stops *k and *g palatalized and became true palatals /c/ and /ɟ/, while *j was simply dropped after other stops.

The second palatalization occured after high vowels were lost, creating new C+j clusters. This time the alveolar stops *t and *d palatalized into alveolo-palatal affricates that later simplify to fricatives /ɕ/ and /ʑ/.

The problem:

I want more alveolo-palatals.

Proposed solution:

have the velars palatalize in both cases.

The question:

How should I go about it?

Should I go with option 1:

1st palatalization: *kj *gj > c ɟ

2nd palatalization: *kj, *tj, *gj, *dj > ɕ, ʑ

Or 2:

1st palatalization: *kj, *gj > *tɕ *dʑ > ɕ ʑ

2nd palatalization: *kj, *tj, *gj, *dj > c, ɕ, ɟ, ʑ

currently I'm leaning towards option 1, because it feels more right, but option 2 has more merging and complication potential which I do enjoy.

[mewling-central]

Here's an evolution history of the word 'Mīriñëu' (Ancestral Language)

[ˌaˈmi.raˌi.ɲaˌɲa.o] => Proto-Ancestral term, compound of [ˌaˈmi.raˌi] "Homeland of the Proto-Ancestral people" and [ˌɲaˈɲa.o] "language, speech".

[ˌa.ˈmi.ˌrae̯.ɲa.ˌɲau̯] => vowels in open syllables preceeding vowels merge, forming diphthongs.

[ˌa.ˈmi.rae̯.ɲa.ˌɲau̯] => secondary stress is lost in syllables preceeding a stressed syllable.

[ˌa.ˈmi.raːɲ.ˌɲau̯] => short vowels lost and diphthongs shorten in unstressed syllables. Pronunciation remains the same until Old Ancestral.

[a.ˈmiː.raːɲ.ˌɲau̯] => short vowels lengthen when preceeding voiced obstruents.

[æ.ˈmiː.raːɲ.ˌɲæu̯] => short vowels become lax in quality.

[æ.ˈmiː.raɲ.ˌɲæu̯] => long vowels shorten preceeding coda consonants.

[ə.ˈmɪː.rəɲ.ˌɲæu̯] => tense vowels become lax, then unstressed short vowels are lost and unstressed long vowels/diphthongs are shortened.

[ˈmɪː.rəɲ.ˌɲæu̯] => /ə/ is lost word-initially and finally.

[ˈmɪː.rɪːɲ.ˌɲæu̯] => /ə/ assimilates to bordering primary syllable vowel, or secondary if not near a primary. Pronunciation remains the same until Middle Ancestral.

[ˈmɪː.ɾɪːɲ.ˌɲæu̯] => /r/ becomes /ɾ/ intervocalically.

[ˈmɪː².ɾɪːɲ².ˌɲæu̯²] => syllables with long vowels, diphthongs, or coda consonants become heavy.]

[ˈmɪː́².ɾɪː́ɲ².ˌɲǽu̯²] => heavy syllables develop high pitch.

[ˈmɪː́².ɾɪː́².ˌɲǽu̯²] => consonants degeminate.

[ˈmɪː́.ɾɪː́.ˌɲǽu̯] => heavy syllable distinction is lost.

[ˈmɪ́.ɾɪ́.ˌɲǽu̯] => long vowels shorten.

[ˈmɪ́.ɾɪ.ˌɲǽu̯] => high pitch vowels in unstressed syllables become unpitched. This is the final pronunciation into the Classical Ancestral period.

[sadpurpleblood]

update: still feeling like shit gonna try and conlang a little anyways cuz ill probably feel worse if i dont get shit done lol not relevant to the thing im actually conlanging rn, but maybe for a descendant of it, do yall think this sequence of consonant evolution seems plausible?: tu > tˠu > t͡ku > |u tuj > tˠuj > tˠi > t͡ki > |i ti > ti those are clicks if thats not obvious. thanks in advance conlangers. love you all

[dedalvs]

Is there a way to evolve ejectives from plain consonants?

The most obvious way is voiceless stop + glottal stop sequences. These can happen easily enough word-internally:

*/ap.ʔa/ > [a.p’a]

Another way would be syncope leading to the consonant cluster leading to the ejective:

*/pa.ʔa.na/ > */pǝ.ʔa.na/ > */pʔa.na/ > [p’a.na]

Those are a couple ways off the top of my head. I’m also fairly certain I’ve heard of geminates becoming ejectives, though I don’t have a specific source on that (hopefully if I’m right someone can point to an example—or if someone knows more about it, they can chime in and say it’s unattested and/or unlikely). Another way is simply through contact. This isn’t the first thing that pops to mind for conlangers, since it presupposes the creation of another language, but it happens a lot in the real world. There are languages that borrow entire series of consonants from other languages (cf. retroflex consonants coming to the Indic languages from Dravidian). Perhaps a language that has only plain stops simply borrows an ejective series from a neighboring language through heavy contact and lots of borrowings. Granted, that’s not evolving ejectives from plain stops, but it is adding them to a language that had none before.

Also, just for anyone else reading this thread, you can absolutely have a series of ejective stops going as far back as your language goes. It’s more common to go from ejectives to non-ejectives than the other way around.

[paizau]

sound change is funny in how many possabilities there are for how things change. Take for example a basic distinction between singleton and geminate consonants: C vs Cː, and the dummy roots *apa vs *appa

In one language, following a development similar to Hebrew, intervocalic singleton stops can lenite to their equivilant voiceless fricative, and geminates shorten:

*apa *appa > *afa *appa > /afa/ /apa/

In another, following a development similar to a fee New Caledonian langauges, geminates become aspirated, and then become voiceless fricatives, while simple stops don't change:

*apa *appa > *apa *apʰa > /apa/ /afa/

literally the opposite reflects. that can be fun to have two different dialects/daughter languages do that

[paizau]

That moment when a root is completely eroded through sound change and all that's left is the inflectional ending:

*rilo-ntad > *rilwonted > *riwonde > *rwōnde > *wōnde > /vo.ⁿde/ - /∅/ + /vo.ⁿde/

compare

*kato-ntad > *ketwonted > *kʰetʰwonde > /kʰe.tʰvo.ⁿde/ - /kʰetʰ-/ + /vo.ⁿde/

[paizau]

I want to get /f/ from some where but don't want to redo /p/ > /f/ because I already did that for Kshafa :(

Oh wait! it just so happens that I have this inconvenient /ts/ that really messes with how my consonant table looks, especially after adding the velarization contrast! so let me just-

/ts/ => /θ/ => /f/

ahh... much better :))

[dedalvs]

If a person wanted to end up with an isolating language, what sort of features would need to be in the proto-language (lets say on a timescale of 1000 years) to end up at an isolating language naturalistically?

It’s really not hard to get an isolating language out of anything. The more morphologically complex it is, though, the longer you’ll need (thinking something like Inuktitut). Basically, though, morphological exponence, whatever it is, loves to stick to the edges of words. This is overwhelmingly the case in all languages. You know what else likes to happen at the edges of words? Phonological erosion. Add those two things together and you get an increasing dependence on analytical constructions. If you lose some distinctions and your language can get by, then it doesn’t really matter, but there are others where it becomes necessary to pull out some sort of auxiliary or particle to have things make sense, and the more that happens, the more speakers get used to it, and the more likely that strategy is to get employed again.

Also, though, this stuff is cyclic—and the classifications crucially depend on your definition of what a “word” is. This is not a trivial distinction. If you look at modern French, it can be quite isolating, if you only listen to the way it’s spoken, and you adhere to traditional word boundaries. If you look at the way things are spelled, it looks more inflectional. On the other hand, if you try to break things up phonologically, these things that sound like words are actually different bits all smushed together—things that, in writing, are separate words, but are actually perhaps better analyzed as single words. If you just do it by the sound of it, modern French is actually quite agglutinative. It’s only history that makes us think of it as a formerly inflectional largely analytical language.

But yeah, as for early features, it really does not matter. You give me anything and I can make an isolating language out of it—or an agglutinating language. It’s not too hard, if you’ve got plenty of time (both in the real world and the fictional world of your conlang).