Random phonology/phonemic inventory thread [2011–2018]

[DesEsseintes]

@Sangfroidish: I like your stuff. There is sth wonderfully weird about Vorinthel. I wish I could add something constructive like shimobaatar but it seems he covered pretty much everything.

---

Mini-idea for sth simple that I had while on holiday, but probably won't do anything with.

/m n/ m n
/b t d k g h/ b t d k g h
/v s/ v s
/l/ l

/a e i o/ a e ı o
/ao̯ ae̯ e̯o eɪ̯ o̯e ʊ̯i/ ao ae eo eı oe uı

Nothing original going on here but the diphthongs are cute.

Sample:
Nanuımıtonı nımo onoeveneda egaeleıbono.

[Man in Space]

- */ɑ̃ õ ũ/ > /a ø y/ in some environments, /ɑn on un/ in others, not sure which yet

Maybe /a ø y/ around coronals, /j/, and front vowels, and /ɑn on un/ elsewhere?

Or */ɑ̃ õ ũ/ > /ɑn on un/ before a stop and > /a ø y/ elsewhere.

Something gleb spit out that I think is cool:

/m n͡m n/
/p pʰ b t͜p t͜pʰ d͜b t tʰ d k kʰ g ʔ/
/t͜s t͜sʰ d͜z/
/t͜ɬ t͜ɬʰ d͜ɮ/
/f v s͜f z͜v z v x ɣ/
/j/

/u ɔ a ɛ i/

(C)V(ʔ)({N d j})({N S(! g) t͜s(ʰ) F[- voice]})

[zyma]

the diphthongs are cute.

I also like the fact that there are only voiced labials, especially since that leaves /v/ as the only fricative besides /s/.

Or */ɑ̃ õ ũ/ > /ɑn on un/ before a stop and > /a ø y/ elsewhere.

Another good possibility. I might suggest generalizing it a bit to */ɑ̃ õ ũ/ > /ɑn on un/ word-finally and before consonants, and > /a ø y/ elsewhere.

[cedh]

- */pˠ fˠ mˠ/ > /kʷ xʷ ŋʷ/ disclaimer: I have no idea whether random inversion of primary and secondary articulations is a thing

Although I could be wrong, I don't think the random, sudden inversion of primary and secondary articulations is a valid/naturalistic sound change.

I think this type of change is quite naturalistic and plausible. All you need is a little bit of imprecision in tongue and lip movement, which would at first assimilate the manner of the secondary articulation to the manner of the primary one, and then weaken the labial component of the sound:
[pˠ fˠ mˠ] > [k͡p x͡f ŋ͡m> [kʷ xʷ ŋʷ]

(Note that at least /k͡p ŋ͡m/ are well-attested as phonemes in various languages, mostly in Africa.)

[Sangfroidish]

- */pˠ fˠ mˠ/ > /kʷ xʷ ŋʷ/ disclaimer: I have no idea whether random inversion of primary and secondary articulations is a thing

Although I could be wrong, I don't think the random, sudden inversion of primary and secondary articulations is a valid/naturalistic sound change. However, this doesn't mean that velarized labials > labialized velars is impossible.

You could have the velarization "drag" the labials farther back in the mouth, making them velars, but then have the velars become labialized due to the fact that they were just labials. Although that still sounds like a stretch to me.

I think any way you go about this, you're going to have to do some diachronic gymnastics. Perhaps for a short period of time between the protolang and modern Vorinthel, all vowels would become unrounded, */i u ũ e o õ æ ɑ ɑ̃/ > */i ɯ ɯ̃ e ɤ ɤ̃ æ ɑ ɑ̃/. After labials and velarized labials, however, these vowels become semi-rounded, */iᵝ ɯᵝ ɯ̃ᵝ eᵝ ɤᵝ ɤ̃ᵝ æᵝ ɑᵝ ɑ̃ᵝ/. A little after this unrounding takes place, the velarized labials become true velars. For some reason, perhaps it's the fact that they were recently labials, the velars that came from velarized labials become pick up labialization from the following vowels. After this, all high and mid back vowels become rounded again, since that seems to be somewhat of a default for that kind of vowel.

Just some ideas. Like I said, diachronic gymnastics.

Yeah, didn't think it'd be quite that simple.

I think this type of change is quite naturalistic and plausible. All you need is a little bit of imprecision in tongue and lip movement, which would at first assimilate the manner of the secondary articulation to the manner of the primary one, and then weaken the labial component of the sound:
[pˠ fˠ mˠ] > [k͡p x͡f ŋ͡m> [kʷ xʷ ŋʷ]

(Note that at least /k͡p ŋ͡m/ are well-attested as phonemes in various languages, mostly in Africa.)

Kinda liking this idea though~

- */z/ > /s/

I could be overlooking something, but doesn't modern Vorinthel lack /s/? What happened to it?

I, er, forgot to type it. Oops.

- elision of */h/; phonemic voiceless sonorants arise from removal of */h/ from the sequences */hm hmˠ hn hŋ hl hɾ/ [hm̥ hm̥ˠ hn̥ hŋ̊ hl̥ hɾ̥]

No voiceless /j/?

*/j/ doesn't exist in the protolang, so it doesn't come from elision of /hj/, it arises from the backing and lenition of /ʃ ʒ/. I guess you could say /ʃ/ should produce /j̊/ and /ʒ/ should produce /j/, but I'm kinda not a fan of having /j̊/ in the modern phonology. Would it be too hand-wavy to say the speakers were crap at distinguishing them and they merged almost right away while the voiceless nasals and liquids magically remained separate? Or something?

When you say you'd rather lose them, do you mean you don't want them in modern Vorinthel, or am I misunderstanding? If you don't want them in modern Vorinthel, then simply don't bother accounting for them in the protolang, and take them out. If you have them in Vorinthel vocabulary, devoice the fricatives and make /ə/ into /ɛ/.

If I am misunderstanding, then I believe it would be very easy to account for the appearance of /v ð ɣ ɣʷ/. Add /b bˠ d g/ to the protolang and have them lenite at around the same time as the aspirated stops. If you want to, you could add /ɢ/ for symmetry, and then have it merge with /g/, or you could just leave it out, as /q/ without /ɢ/ is quite common.

As for /ə/, maybe have it as an epenthetic vowel appearing to split up clusters that, while "legal" in the protolang, came to be considered phonotactic violations in modern Vorinthel. This might not work, since I don't know Vorinthel phonotactics and the places in which /ə/ occurs in modern vocabulary. Syllabic consonant > /ə/ would be my other recommendation, especially since you already have /ɾ̩/ > /ɑr/.

I meant I want them gone (or at least changed heavily) in the protolang but present in the modern form. Sorry, I'm making everything awkward by working backwards, heh.
I was thinking more of throwing in an extra vowel that elides to build clusters rather than developing one to break them up, but I guess there's no reason not to do the reverse; Modern Vorinthel is the one I want to look/sound nice, after all, the protolang is just there for background and deriving sisterlangs, it can be as ugly as it wants Plus Vorinthel never allows clusters of more than two consonants anyway, so.

/ə/ doesn't come up much anyway, I could pretty easily jiggle things around to restrict it to places where historically overlarge clusters stood. Except the fact it shows up as a vowel ending, but I'll work round that somehow.

and I'd like to find a way to eradicate /ŋ/ from the protolang too if possible.

Nothing even remotely easy comes to mind that would achieve that goal and let you keep /ŋ̊ ŋ̊ʷ ŋ ŋʷ/ in modern Vorinthel.

Fair enough. I mostly wanted it gone to allow for deriving other daughters that lacked it,, but that could easily be done by dropping it or merging it to /n/ or something.

It looks good so far!

@Sangfroidish: I like your stuff. There is sth wonderfully weird about Vorinthel. I wish I could add something constructive like shimobaatar but it seems he covered pretty much everything.

Thank you!

[zyma]

I think this type of change is quite naturalistic and plausible. All you need is a little bit of imprecision in tongue and lip movement, which would at first assimilate the manner of the secondary articulation to the manner of the primary one, and then weaken the labial component of the sound:
[pˠ fˠ mˠ] > [k͡p x͡f ŋ͡m> [kʷ xʷ ŋʷ]

If you say so. Is it attested?

(Note that at least /k͡p ŋ͡m/ are well-attested as phonemes in various languages, mostly in Africa.)

I think we're all well aware of that. If memory serves, they're also found in quite a few languages of Oceana, although perhaps not as frequently there as in Africa.

*/j/ doesn't exist in the protolang, so it doesn't come from elision of /hj/, it arises from the backing and lenition of /ʃ ʒ/. I guess you could say /ʃ/ should produce /j̊/ and /ʒ/ should produce /j/, but I'm kinda not a fan of having /j̊/ in the modern phonology. Would it be too hand-wavy to say the speakers were crap at distinguishing them and they merged almost right away while the voiceless nasals and liquids magically remained separate? Or something?

Ahh, OK. It seems I overlooked the lack of /j/ in the protolang. And it's perfectly OK to not make a voicing distinction in /j/, even if all the other sonorants have one.

I meant I want them gone (or at least changed heavily) in the protolang but present in the modern form. Sorry, I'm making everything awkward by working backwards, heh.

So, was the voiced stop idea along the lines of what you're looking for? Or do you want the protolang forms to be even more "heavily changed"?

Fair enough. I mostly wanted it gone to allow for deriving other daughters that lacked it,, but that could easily be done by dropping it or merging it to /n/ or something.

Oh, yeah, getting rid of the velar nasal should be easy.

[Sangfroidish]

The voiced stops will do just fine. One of the things I was thinking of going for was to have the protolang be pretty plosive-heavy and light on fricatives and then turn it around, so not quite sure why it didn't occur to me to put them in to begin with really.

[zyma]

The voiced stops will do just fine. One of the things I was thinking of going for was to have the protolang be pretty plosive-heavy and light on fricatives and then turn it around, so not quite sure why it didn't occur to me to put them in to begin with really.

Awesome! And you can't go wrong with the good old plosive-fricative flip.

[Dezinaa]

I'm thinking about removing /ɸ ð/ and replacing /ʉ ə/ with /ui̯ eo̯/ Would that be realistic (having /ui̯/ and /eo̯/ without more common diphthongs)?

/m n ŋ/ <m n ng>
/b t k/ <b t k>
/mb nd ŋg/ <mb nd ngg>
/ɸ β ð s z ɣ h/ <f v th s z gh h>
/l j w/ <l y w>
/ǀ ǃ ǁ ᶢǀ ᶢǃ ᶢǁ ᵑǀ ᵑǃ ᵑǁ/ <c q x gc gq gx nc nq nx>
/i ʉ u e ə o a aɪ̯/ <i ui u e eo o a ai>

[zyma]

I'm thinking about removing /ɸ ð/ and replacing /ʉ ə/ with /ui̯ eo̯/ Would that be realistic (having /ui̯/ and /eo̯/ without more common diphthongs)?

/m n ŋ/ <m n ng>
/b t k/ <b t k>
/mb nd ŋg/ <mb nd ngg>
/ɸ β ð s z ɣ h/ <f v th s z gh h>
/l j w/ <l y w>
/ǀ ǃ ǁ ᶢǀ ᶢǃ ᶢǁ ᵑǀ ᵑǃ ᵑǁ/ <c q x gc gq gx nc nq nx>
/i ʉ u e ə o a aɪ̯/ <i ui u e eo o a ai>

I'd say remove the two fricatives and make /ui̯ eo̯/ allophones of /ʉ ə/.

[stonyhonu]

Tried to make a minimalistic phonology inspired by polynesian languages and somewhat japanese.
/m n ŋ/ <m n g>
/p t k ʔ/ <p t k '>
/s h/ <s h>
/l/ <l>
/a ɛ i o/ <a e i o>
/ɐ̃ ẽ ɨ̃ ʊ̃/ <an en in on>

CV only. Roots can be up to (CV)(CV)CVCV with derived roots.
Glottal stop only written after a nasal vowel to disambiguate.
Stress always on penultimate syllable. Secondary stress on preantepenultimate syllable.
o, ɛ [+pre-tonic] > u, e
n, t, s, h, l > ɲ, t͡ʃ, ʃ, ç~x, ʎ / _i, _ɨ̃

maetelokan'a [maʔeˌtɛluˈkɐ̃ʔa]
pegatasii [peˌŋataˈʃiʔi]
kohan'osele [kuˌhɐ̃ʔuˈsɛlɛ]

So, some questions:
1. Should I add /f/, or maybe /v/, or neither?
2. I was thinking about adding an aspirated stop series and changing the plain series to ejectives. I like this idea but for the fact that I'm not sure I'm sold on <ph th kh> for the apirated stops, and I know I don't want <p' t' k'> for the ejectives. So without a solid romanization* I'm not sure I want to do this. Maybe I will add <ph th kh>, and change the non-aspirates, just to differentiate it more from polynesian languages. What do you think?
3. If no to question #2 (or even if yes), should I add something like coarticulated stops (maybe kp & ŋm) or maybe another idea I haven't thought of that could differentiate it more from polynesian languages, but not really lose too much of the style; for instance, I don't want to add many more fricatives.

*Speaking of romanization, I'm trying to come up with a cool syllabary but the first draft was a katakana ripoff. This is definitely the more difficult part for me.

[zyma]

I like how nasalization changes vowel quality, the fact that there's secondary stress, and the allophony.

I'd say no to /f/ and /v/.

I personally would leave it as it is, but if you really want aspirated vs. ejectives, but you don't like <ph th kh p' t' k'>, you could go with: /pʰ tʰ kʰ p' t' k'/ <p t k b d c> or /pʰ tʰ kʰ p' t' k'/ <b d c p t k>. Since you already have <g> for the velar nasal, I think <c> could work as the "partner" for <k>. Just an idea.

[DesEsseintes]

An idea for a vowel inventory:

/a i u ɯ/

Add length and closing diphthongs /ai au aɯ/ [aɪ̯ aʊ̯ aɯ̯̽], and this might be quite fun.

Edit: In the last hour or so, this has grown to

/m n ŋ~ɴ/
/p t k~q/
/s x~h/
/l/
/a i u ɯ/

Expect abundant allophony.

[zyma]

An idea for a vowel inventory:

/a i u ɯ/

Add length and closing diphthongs /ai au aɯ/ [aɪ̯ aʊ̯ aɯ̯̽], and this might be quite fun.

Edit: In the last hour or so, this has grown to

/m n ŋ~ɴ/
/p t k~q/
/s x~h/
/l/
/a i u ɯ/

Expect abundant allophony.

Looks nice so far. Any plans regarding romanization? I'm curious as to how you'll represent the velar/uvular nasal and the close back unrounded vowel.

[DesEsseintes]

Looks nice so far. Any plans regarding romanization? I'm curious as to how you'll represent the velar/uvular nasal and the close back unrounded vowel.

Now I'm thinking the phonology consists of four different kinds of phonemes:
a) stops /p t k~q ʔ/
b) glides /j w ɰ/
c) colorations /ɴ ʟ ʜ/
d) the vowel /a/

Many morphemes are "areal" infixes of glides, n-coloration and l-coloration.

I need a lot of allophones for each phoneme.

Here are some sketches, but I'm not really sure what I'm doing. Furthermore, there are inconsistencies within just these few examples, because they're sketches.

jʜ-t-twt → ıþþtut
j-ʜj-jt → eıhhyıt

ja-ɴw-aɴ → [jɒw̃an] yanwan
jw-ɴw-aɴ → [joʊ̯w̃ːan] yunnwan
at-jw-ɴw-aɴ → [d͡ʑoʊ̯w̃ːan]* dzyounwan
at-jw-ɴw-jaɴ → [d͡ʑoʊ̯mːjan] dzyoummyan
tw-ɴɰ-jkkja → [tuɴːɯ̯̽eɟːæ] tunveıggya
tw-ɴɰ-jʜkkja → [tuɴːɯ̯̽eɕːcæ] tunveıśśkya
*the idea being that the initial a elides but leaves the consonant voiced

So that's where it's at now...

[zyma]

Looks nice so far. Any plans regarding romanization? I'm curious as to how you'll represent the velar/uvular nasal and the close back unrounded vowel.

Now I'm thinking the phonology consists of four different kinds of phonemes:
a) stops /p t k~q ʔ/
b) glides /j w ɰ/
c) colorations /ɴ ʟ ʜ/
d) the vowel /a/

Many morphemes are "areal" infixes of glides, n-coloration and l-coloration.

I need a lot of allophones for each phoneme.

Here are some sketches, but I'm not really sure what I'm doing. Furthermore, there are inconsistencies within just these few examples, because they're sketches.

jʜ-t-twt → ıþþtut
j-ʜj-jt → eıhhyıt

ja-ɴw-aɴ → [jɒw̃an] yanwan
jw-ɴw-aɴ → [joʊ̯w̃ːan] yunnwan
at-jw-ɴw-aɴ → [d͡ʑoʊ̯w̃ːan]* dzyounwan
at-jw-ɴw-jaɴ → [d͡ʑoʊ̯mːjan] dzyoummyan
tw-ɴɰ-jkkja → [tuɴːɯ̯̽eɟːæ] tunveıggya
tw-ɴɰ-jʜkkja → [tuɴːɯ̯̽eɕːcæ] tunveıśśkya
*the idea being that the initial a elides but leaves the consonant voiced

So that's where it's at now...

It's quite different from the last version, at least from my perspective, but I think I might like this even more! I can't claim to understand all of what's going on, but it sure looks awesome!

[Shemtov]

Idea that came to me; won't do anything with:
/p t k q ʔ/ <p t k q '>
/ m n/ <m n>
/f v s ɬ X h/ <f v s ll kh h>
/r ʀ/ <rr r>
/l/ <l>


/i ɯ u/ <i ui u>
/e ɤ o/ <e oe o>
/a/ <a>

/ai oi au ou aɯ ɤi ɤɯ ɤo ɯu uɯ oɤ/ <ai oi au ou aui oei oeu eo iu uu oo>

[zyma]

This phonology's consonant inventory was heavily inspired by Japanese, but I was aiming for a more Turkic/Uralic feel with the vowels.

/p b t d k g/ p b t d k g
/ɸ s z~d͡z ɕ ʑ~d͡ʑ h/ f s z ṣ j h
/t͡s t͡ɕ/ c ṭ
/m n ɴ/ m n ṇ
/j w/ y w
/ɺ/ r

/i y e ø æ/ i ü e ö ä
/ɯ u ɤ o ɑ ɒ/ ı u ɛ o a å

/ː/ Gemination is represented by doubling a grapheme. Both consonants and vowels can be geminated.

There will probably be some sort of vowel harmony. Maybe i y e ø æ vs ɯ u ɤ o ɑ ɒ, or i e æ ɯ ɤ ɑ vs y ø u o ɒ, or i y ɯ u vs e ø ɤ o vs æ ɑ ɒ…

Also, I hope to use some sort of consonant gradation/mutation. Probably:

/p b t d k g/ > /b w d z g h/
/pː bː tː dː kː gː/ > /p b t d k g/
/ɸ s z~d͡z ɕ ʑ~d͡ʑ h/ > /w z~d͡z r ʑ~d͡ʑ j Ø/
/ɸː sː zː~d͡zː ɕː ʑː~d͡ʑː hː/ > /ɸ s z~d͡z ɕ ʑ~d͡ʑ h/
/t͡s t͡ɕ/ > /z~d͡z ʑ~d͡ʑ/
/t͡sː t͡ɕː/ > /t͡s t͡ɕ/
/m n ɴ/ > /pː tː kː/
/mː nː ɴː/ > /m n ɴ/
/j w/ > /Ø[+front] Ø[+back]/
/jː wː/ > /j w/
/ɺ/ > /Ø/
/ɺː/ > /ɺ/

I might add more tomorrow. There are probably tons of typos in this since I'm so tired.

[zyma]

I'm not sure what I want to do with this, but I like it. I might make the language polysynthetic, or at least something with noun incorporation and possessive suffixes.

/t k ʔ/ t c q
/s̻~z̻ s̺~z̺ h/ z s h
/m n/ m n
/j w ɰ/ y w g
/ⱱ ɽ/ v r

/i e/ i e
/u o ɑ/ u o a

/aɪ̯/ æ

I'm not sure about the coronal fricative distinction. I might keep it laminal vs. apical, or I might change it to dental vs. alveolar. I plan on having vowel harmony (front vs. back). The diphthong æ will pattern with the front vowels.

I kind of want to do something similar to Japanese and have /ʔ h/ become [p ɸ] before /u o/. I might represent these allophones in the orthography as p f, but I'm not sure.

I also want to have /n/ become [ŋ] word-finally, and have c /k/ palatalize to [t͡ʃ] before /i e/.

Phonotactics may be either (C)V, (C)V(n), or something similar.

Examples:

zoga /s̻oɰɑ/ deity, spirit
mæ /maɪ̯/ to see, to look at, to observe
imi /imi/ back

[DesEsseintes]

inventory

I like this. Personally, I would add whatever distinction you go for in the sibilants in the stops too, but that's because I can't control myself.

I plan on having vowel harmony (front vs. back). The diphthong æ will pattern with the front vowels.

How about æ being allowed as a transition phoneme from back to front harmony? As an example tokonæqite could be a word? Just an idea.

I kind of want to do something similar to Japanese and have /ʔ h/ become [p ɸ] before /u o/. I might represent these allophones in the orthography as p f, but I'm not sure.

Love it! The reasoning behind it could be that [ʔ] was originally [kʷ] that backed to [q] then debuccalised to [ʔ] except before back roundeds where it became labial instead.

Or that's just what I thought up when I saw this.