Random phonology/phonemic inventory thread

[eldin raigmore]

@ Omzinesý :
Shouldn’t some of your latest post be continued on the
Random Ideas: Morphosyntax
https://cbbforum.com/viewtopic.php?t=712
topic?

[Porphyrogenitos]

Inspired by Indic schwa-deletion, and designed to be underlyingly CV and writable with a pure CV syllabary.

/m n/ m n
/p t k/ p t k
/ɾ s/ r s

/i u/ i u
/ə/ e
/a/ a

/ə/ can be characterized as short or "weak", while /i a u/ are long or perhaps "strong". It is possible that /ə a/ were originally a short-long pair.

In any case, /ə/ is vulnerable to deletion, while /i a u/ are not. Schwa-deletion and stress rules may vary from dialect to dialect, but this is the system I am proposing provisionally:

Stress is on the antepenultimate syllable, unless it is short and the preceding syllable is long, in which case stress falls on the preantepenult. (On disyllables stress is on the first syllable.) Secondary stress is assigned both leftward and rightward from the stressed syllable to every other syllable, excepting the final syllable.

Final /ə/ is always deleted. This rule is probably constant across all dialects.

Schwa in an unstressed syllable is deleted whenever it would result in a word-medial CC cluster. So, for instance:

/patəkaɾisə/ becomes [patˈkaɾis]
but /təkaɾisə/ becomes [təˈkaɾis],with the first schwa not being deleted, because it would create an impermissible word-initial cluster

Final schwa may thus resurface when a sufficient number of affixes are added, e.g.

/pikə/ → [pik]
/pikəma/ → [ˈpikma]
/pikəmasa/ → [ˈpikˌmasa]
/pikəmasami/ → [ˌpikˈmasami]
/pikəmasamiɾi/ → [piˌkəmaˈsamiɾi]

Although I'm wondering if there might not just develop a rule that secondary stress on schwa also moves back to an eligible long vowel.

There may be additional rules about assimilation of consonants and such, but these probably vary by dialect.

One additional feature I am considering is neutralization of surface coda nasals to nasalization of the preceding vowel, e.g. both /kamə/ and /kanə/ are [kã], /taməki/ and /tanəki/ are both [tãki], etc.

The language can conveniently be represented with a syllabary of just 32 symbols. I have also been thinking about an anusvara-like segment - that is, phonemic nasalization of the preceding vowel, or perhaps some form of syllabic nasal - which may block schwa-deletion or have other effects. I'm not sure whether it would be represented by a 33rd symbol or as a diacritic.

[Shemtov]

Once I more fully flesh out gǂàamą́ 3oį-ʔą̀rhokìi, I want to redo Waanian, a language that has gone through multiple versions. It is based on Austronesian, but for the new version I want to incorperate Mayan influences, and thus I have produced the following phonology; I may reapropiate this phonology to be Proto-Waanic, and diachronically derive the phonology of the prestige dialect of what I am temporarily calling "Western South-Peninsular Southwest Waanic", but am not sure if this is the protolang, or the language itself; to that end, phonemes I want to include only if this is the protolang will be in parenthesis:

/p b ɓ t d t' cç ɟʝ cç' k g k' q (ɢ) q' ʔ/
/m n ɲ ŋ/
/s ç h/
/r ʀ/
(/ɾ/)
/l/
/j w/

/i ɨ u a/
/i: u: a:/
/ai au/
(/a:i a:u/)


Phontactics: (C)V(C)
Permitted finals:
/p b t d k g q (ɢ) s h m n ɲ ŋ r ʀ l j w/

[Shemtov]

So, I've decided that is the Proto-Lang, and I've evolved a Waanic language, Zeigouŋdeizese, which I will work on when I've more fully flesh out gǂàamą́ 3oį-ʔą̀rhokì:
/p b~β ɓ t d~ð t' t͡ʃ d͡ʒ t͡ʃ' k g~ʁ k' ʔ/ <p b b' t d t' c ǰ c' k g k' '>
/m n ɲ ŋ/ <m n ñ ŋ>
/θ s ɬ ʃ ʒ h/ <z s ł x j h>
/r / < ř>
/ɾ/ <r>
/l/ <l>
/w j/ <w y>

/i ɛ u ɔ a/ <i e u o a>
/i: u: a:/ <ii uu aa>
/ai au ɛi ɔu ɛu/ <ai au ei ou eu>

Phontactics: (C)V(C)
Permitted finals:
/p b t d k g h m n ɲ ŋ r l j w/
The voiced stops take their fricative allophone when syllable-final

[eldin raigmore]

@Shemtov :
Whence cometh the name “Waanic”?
It seems somewhat similar to the name of an Australian Aboriginal language that is unfortunately recently extinct.

[thethief3]

*p *t *tʲ *k
*t͡s *t͡sʲ
*(f) *s *sʲ *ʃ
*ɬ *lʲ *ɾ
*m *n *nʲ

*i *u
*e *o
*ɛ *ɔ
*a

All vowels distinguished 3 degrees of length although mid long high vowels only occurred in Amarin loanwords. In unstressed (non-initial) syllables they only distinguished two.

*p and *f are mainly in complementary distribution with *f occurring word initially and *p otherwise. This is the case except in some onomatopoeia.

CVC syllable structure maximally

*ɛ:mɛɾin “Amarin”, *kutɛna “person”, *sɔli “fish”, *nɛrɛi “sky”, *t͡sʲuna “soul”, *ta:pa “idol”, *tɛ:nɛ “sea” *sɛ::wi “skunk”, *enta “to dance”, *kuɾi “shaman”, *fat͡san “doctor, *folʲi “city”, *pɔkɔ “human”, *ʃensiwi “ice”, *tuki “sound”, *ta:jalʲa “salmon”, *miʲta “seamstress”, *ɛ:na “daughter”, *ɬuwɛn “son”, *kappa “ship”, *t͡sʲana “to go”, *nʲenʲe “wing”, *kilsi “star”, *mira “soul”, *a:na “mother”, *na:kasi “father”, *ku::ma “to beget”, *ka::si “snake”, *ko::na “blue”, *su::pi “to exceed”, *tʲi::nʲsʲi “drum”

[Shemtov]

@Shemtov :
Whence cometh the name “Waanic”?
It seems somewhat similar to the name of an Australian Aboriginal language that is unfortunately recently extinct.

Located about 4KM off the Northeast coast of Fuhe, Waan (Fuhek: Vaan) is the Newfoundland-sized Island the Waanic languages are native too, coming from PWaanic *waan, which originally meant "Dry land", but the meaning shifted in Western Waanic to be a proper name, but the original meaning is preserved in Zeigouŋdeizese in various religious terms (i.e. "Great Waan" to mean a "The sentient vibration of the World Soul", "Ground Waan", for "The Underworld Paradise", and "Waandeiza" "Waan City" or "Waan Place" as "Class of certain sacred sites and mythical places")

[DV82LECM]

Removed.

[Shemtov]

So, I've decided that is the Proto-Lang, and I've evolved a Waanic language, Zeigouŋdeizese, which I will work on when I've more fully flesh out gǂàamą́ 3oį-ʔą̀rhokì:

Slightly revised inventory:
/p b~β ɓ t d~ð t' t͡ʃ d͡ʒ t͡ʃ' k g~ʁ k' ʔ/ <p b b' t d t' c c' j k g k' '>
/m n ɲ ŋ/ <m n ñ ŋ>
/θ s ɬ ʐ h/ <z s ł x h>
/r / < ř>
/ɾ/ <r>
/l/ <l>
/w j/ <w y>

/i e u o a/ <i e u o a>
/i: e: u: a:/ <ii ei uu o aa>
/ai au ɛu/ <ai a eu>

Phontactics: (C)V(C)
Permitted finals:
/p b t d k g h m n ŋ r l j w/
The voiced stops take their fricative allophone when syllable-final

[Porphyrogenitos]

The latest of my six-phoneme absolutely minimal phoneme inventories, this time missing velars and fricatives:

/m n/ m n
/p t/ p t
/β̞ ɾ/ b r

/β̞ ɾ/ may be [ b d] utterance-initially.

/t/ is [tʃ] before /i/.

/a i o/ a i o

/o/ instead of /u/ for an aesthetic that's a bit... softer? I guess. Thinking of Piraha and some NAm languages.

(C)V syllables, naturally. To be represented by a 21-symbol syllabary.

As for suprasegmentals... I like to make these minimalistic languages without tone, and generally without phonemic stress, both for my own ease, and as a bit of a challenge since it makes it even more minimal and reliant on the segmental phonology. But, I admit, it would be kind of aesthetically fitting to have some kind of tone or phonemic stress here. However, for the moment, I will leave it as it is.

A nonsense paragraph generated via Lexifier:

Tati ti mabotapito tamima ta. Ta poni ta tabimata ipira? Bamari ti ta
ma patana. Ti ta mota patitara tama. Ma ni tari rata pita tanaba. Tama
mati matibiri tana amati? Mi ma amata nama batamita amiripi nanita?
Atino tapani tita amati ramapi naba, napa otamarati na ma rata
batabato? Rama mi papino no matamana, ri mani mora. Tamo papitama tapo
na, rana rata tinatamo miti. Mati ta atonanima napa ma. Natimi pa ta
nara tito pami. Mapimotapo tamapipata mata momitama.

[Omzinesý]

Not many ideas for this lang.
I just wanted to have a lang with "nasal affricates", ie. vowel nasalization is phonemic only after a nasal consonant.

p t t͡s k kh
b d d͡z g
f s x
v z
m n ŋ
mṼ nṼ ŋṼ
l j

i: u:
ɪ ʊ
e: o:
ɛ ɔ
æ: ä ɑ:

[DesEsseintes]

/p t t͡ʃ k/ p t ch k
/f s/ f s
/m ɾ w/ m r w

/a e i o/ a e i o

All consonants except /ɾ w/ can occur geminate. Onset clusters include /ps ks/ and some Cw clusters but not much else.

/s/ palatalises to /ʃ/ sh before /i/.

Words cannot start in coronals /t s ɾ/
Words can end in any vowel, or one of /t͡ʃ k s m/

[DV82LECM]

/p t t͡ʃ k/ p t ch k
/f s/ f s
/m ɾ w/ m r w

/a e i o/ a e i o

All consonants except /ɾ w/ can occur geminate. Onset clusters include /ps ks/ and some Cw clusters but not much else.

/s/ palatalises to /ʃ/ sh before /i/.

Words cannot start in coronals /t s ɾ/
Words can end in any vowel, or one of /t͡ʃ k s m/

Nice, but why /ps ks/ at onset, and not /s/?

[DesEsseintes]

/p t t͡ʃ k/ p t ch k
/f s/ f s
/m ɾ w/ m r w

/a e i o/ a e i o

All consonants except /ɾ w/ can occur geminate. Onset clusters include /ps ks/ and some Cw clusters but not much else.

/s/ palatalises to /ʃ/ sh before /i/.

Words cannot start in coronals /t s ɾ/
Words can end in any vowel, or one of /t͡ʃ k s m/

Nice, but why /ps ks/ at onset, and not /s/?

Why indeed.

I haven’t really given it much thought yet, but let us suppose the restriction on no word-initial coronals predates these clusters. The clusters could subsequently have developed in one of several ways.
1) /ps ks/ could have developed from earlier/pj kj/ (as is common for instance in Blackfoot). This would be an elegant solution as the modern language lacks /j/ and this would then have been a part of a general process of depalatalisation.
2) /ps ks/ could result from syncope in word-initial *pVs- *kVs- sequences.

Either way, the occurrence of word-initial ps ks without corresponding initial s seems to me relatively straightforward to justify.

[DV82LECM]

/p t t͡ʃ k/ p t ch k
/f s/ f s
/m ɾ w/ m r w

/a e i o/ a e i o

All consonants except /ɾ w/ can occur geminate. Onset clusters include /ps ks/ and some Cw clusters but not much else.

/s/ palatalises to /ʃ/ sh before /i/.

Words cannot start in coronals /t s ɾ/
Words can end in any vowel, or one of /t͡ʃ k s m/

Nice, but why /ps ks/ at onset, and not /s/?

Why indeed.

I haven’t really given it much thought yet, but let us suppose the restriction on no word-initial coronals predates these clusters. The clusters could subsequently have developed in one of several ways.
1) /ps ks/ could have developed from earlier/pj kj/ (as is common for instance in Blackfoot). This would be an elegant solution as the modern language lacks /j/ and this would then have been a part of a general process of depalatalisation.
2) /ps ks/ could result from syncope in word-initial *pVs- *kVs- sequences.

Either way, the occurrence of word-initial ps ks without corresponding initial s seems to me relatively straightforward to justify.

Fair enough. I simply wanted to see IF you had derived something. Oddly enough, River Skasti has the same arrangement of /s ps ks/, but /s/ is initial. These were /ts pts kts/. I am way too slavish to congruence.

[Porphyrogenitos]

Yet another nasal-less, labial-less, fricative-less inventory.

/d ɖ ɟ g gʷ/ d dd j g gw
/t ʈ c k kʷ ʔ/ t tt c k kw '
/t' ʈ' c' k' kʷ'/ t' tt' c' k' kw'
/l ɾ ɻ j ɰ w/ l r rr y h w
/l̰ ɾ̰ ɻ̰ j̰ ɰ̰ w̰/ l' r' rr' y' h' w'

/a e i o u/ a e i o u

My feeling is that glottalization will be kind of like Arabic pharyngealization in that, in practice, it will be kind of autosegmental and spread across an entire word, so that in a word like k'al, the /l/ will end up being [l̰] even if not indicated in the orthography. In a countervailing manner, there will probably be Grassman's/Dahl's-law-like dissimilation of ejectives (only one ejective stop per word). And as with ancient Indo-Aryan, there will probably be immediate and long-distance retroflex assimilation phenomena.

One idea I have for diachronics is that /ɾ/ is relatively new, originating mainly from recent lenition of /t/ and /d/, with /ɾ̰/ being marginal and mainly originating from long-distance glottalization of /ɾ/.

Oh, and my favored realization for /ɰ̰/ would probably be a creaky-voiced glottal approximant [ʔ̞], although a voiced pharyngeal approximant [ʕ] would be good too.

[DV82LECM]

Yet another nasal-less, labial-less, fricative-less inventory.

/d ɖ ɟ g gʷ/ d dd j g gw
/t ʈ c k kʷ ʔ/ t tt c k kw '
/t' ʈ' c' k' kʷ'/ t' tt' c' k' kw'
/l ɾ ɻ j ɰ w/ l r rr y h w
/l̰ ɾ̰ ɻ̰ j̰ ɰ̰ w̰/ l' r' rr' y' h' w'

/a e i o u/ a e i o u

My feeling is that glottalization will be kind of like Arabic pharyngealization in that, in practice, it will be kind of autosegmental and spread across an entire word, so that in a word like k'al, the /l/ will end up being [l̰] even if not indicated in the orthography. In a countervailing manner, there will probably be Grassman's/Dahl's-law-like dissimilation of ejectives (only one ejective stop per word). And as with ancient Indo-Aryan, there will probably be immediate and long-distance retroflex assimilation phenomena.

One idea I have for diachronics is that /ɾ/ is relatively new, originating mainly from recent lenition of /t/ and /d/, with /ɾ̰/ being marginal and mainly originating from long-distance glottalization of /ɾ/.

Oh, and my favored realization for /ɰ̰/ would probably be a creaky-voiced glottal approximant [ʔ̞], although a voiced pharyngeal approximant [ʕ] would be good too.

Such a Pama-Nyungan feel. If it were, it would be my second favorite to Arrernte.

[VaptuantaDoi]

Yet another nasal-less, labial-less, fricative-less inventory.

/d ɖ ɟ g gʷ/ d dd j g gw
/t ʈ c k kʷ ʔ/ t tt c k kw '
/t' ʈ' c' k' kʷ'/ t' tt' c' k' kw'
/l ɾ ɻ j ɰ w/ l r rr y h w
/l̰ ɾ̰ ɻ̰ j̰ ɰ̰ w̰/ l' r' rr' y' h' w'

/a e i o u/ a e i o u

Beautiful!

My feeling is that glottalization will be kind of like Arabic pharyngealization in that, in practice, it will be kind of autosegmental and spread across an entire word, so that in a word like k'al, the /l/ will end up being [l̰] even if not indicated in the orthography. In a countervailing manner, there will probably be Grassman's/Dahl's-law-like dissimilation of ejectives (only one ejective stop per word). And as with ancient Indo-Aryan, there will probably be immediate and long-distance retroflex assimilation phenomena.

*chef's kiss*

In fact I like this so much I'll make one myself. This takes more Lillooet influence.

/tʰ tʃʰ ʈʂʰ kʰ kʷʰ qʰ ʡ/ ⟨t c c̣ k kʷ q ɂ⟩
/t’ tʃ’ ʈʂ’ k’ kʷ’ q’ ʔ/ ⟨t̓ c̓ c̣̓ k̓ k̓ʷ q̓ ɂ̓⟩
/z l ɭ j w ʁ ʕ/ ⟨z l ḷ y w r g⟩

/e o ɛ₁ ɔ ɛ₂ a/ ⟨e o ẹ ọ a ạ⟩

- /kʰ k’/ are replaced by [tɕʰ tɕ’] ⟨ç ç̓⟩ before /e/
- /e o/ presumably become [​i u] in some environments

All non-dental phonemes are specified [±dark], with harmony applying across a word.
[-dark] /tʃʰ kʰ qʰ tʃ’ k’ q’ l j ʁ e o ɛ₂/
[+dark] /ʈʂʰ kʷʰ ʡ ʈʂ’ kʷ’ ʔ ɭ w ʕ ɛ₁ ɔ a/
neutral: /tʰ t’ z/

Alternatively, a vertical vowel system
[-dark] /i ɛ/ ⟨ı a⟩
[+dark] /u ɑ/ ⟨i ạ⟩

I guess if you analysed [±dark] as suprasegmental you only need
/tʰ tʃʰ kʰ qʰ/
/t’ tʃ’ k’ q’/
/z l j ʁ/
/i ɛ/ or /e o ɛ/

[Porphyrogenitos]

Such a Pama-Nyungan feel. If it were, it would be my second favorite to Arrernte.

Beautiful!

*chef's kiss*

Thank you both! It was inspired by a long-standing curiosity I've had about the possible intersection of the three rare but attested phonological gaps featured here.

In fact I like this so much I'll make one myself. This takes more Lillooet influence.

This is very neat! I definitely like the inclusion of the affricates and do find them more aesthetic (although as DV82LECM noted I guess using palatal stops gives mine an Australian feel). Affricates aren't fricatives of course but I had wondered if, by including them, I was skirting too close to the edges of the limits I had set for myself.

I really put no thought into my vowel system - maybe I ought to have made it /i a u/ and added Semitic or Australian inspired allophonic phenomena.

I suppose I ought to actually develop something with this inventory since it was received so well.

[Omzinesý]

t' k' q'
p t k q ʔ
f θ s ɬ x χ ħ h
ʋ ɹ ɣ ʁ ʕ
m n (ɴ)

Consonant clusters: stop + continuant