Post your conlang's phonology

[Nortaneous]

tɬoman aɬphaβet tɕonɬanŋ

/p t ts tɬ tɕ k ʔ/ <p t z r c k q>
/β ð/ <b d>
/ɸ s ɬ x ħ h/ <f s l j x h>
/m n ŋ/ <m n g>
/a e ə o i y ɯ u ə̃/ <a e y o i w u v>

[CatDoom]

/t k m n r/
/i y e ø ɛ œ æ a ɨ ʉ ɘ ɵ ɜ ɞ ɯ u ɤ o ʌ ɔ ɑ/, C(r)V(C) syllable structure, etc...

Hmm... I know it doesn't make much sense to nitpick a phonology as crazy as this one, but how common is it for a language to have /m/ as its only labial consonant? I know there are a fair number of languages in the Americas with no bilabial consonants at all; maybe you could have something like a truncated version of Rotokas, with /t d~ɾ k ɡ~ɣ n/.

[Nortaneous]

/t k m n r/
/i y e ø ɛ œ æ a ɨ ʉ ɘ ɵ ɜ ɞ ɯ u ɤ o ʌ ɔ ɑ/, C(r)V(C) syllable structure, etc...

Hmm... I know it doesn't make much sense to nitpick a phonology as crazy as this one, but how common is it for a language to have /m/ as its only labial consonant? I know there are a fair number of languages in the Americas with no bilabial consonants at all; maybe you could have something like a truncated version of Rotokas, with /t d~ɾ k ɡ~ɣ n/.

/t k n ŋ h/
/i y ɯ u ɪ ʏ ɷ ʊ e ø ɤ o æ œ ɑ ɒ ɯ˞ u˞ ɤ˞ o˞ ɑ˞/

velars/glottal -> labials before rounded vowels

[cromulant]

/t k m n r/
/i y e ø ɛ œ æ a ɨ ʉ ɘ ɵ ɜ ɞ ɯ u ɤ o ʌ ɔ ɑ/, C(r)V(C) syllable structure, etc...

Hmm... I know it doesn't make much sense to nitpick a phonology as crazy as this one, but how common is it for a language to have /m/ as its only labial consonant?

Cherokee. Hupa, almost/sorta: its labials are /m w/ and [p], sez Wikipedia. So strictly speaking, /m/ is its only phonemic bilabial. And my lang (which I am taking more seriously than the jokelang I should probably view it as) does have [p] and [w] (well, C^w) pop up in certain environments.

----

So then, so then. Here is a vowel chart for easy visual reference:

Code: Select all

i y     ɨ ʉ     ɯ u
e ø     ɘ ɵ     ɤ o
ɛ œ     ɜ ɞ     ʌ ɔ
æ
a               ɑ

Overlaying these 21 basic vowel qualities are distinctions of nasality and length.

And now the consonants:

Code: Select all

     t     k
m    n
     ɾ

(When referring to the tap phoneme, I will simply use regular <r> rather than the IPA symbol).

Syllable shape is (C)(r)V. This includes rrV, where the doubled r is trilled, not tapped as it is when it stands alone.

Null onsets are allowed only in word-initial position. In content words, they are required in that position, and are separated from the final vowel of the preceding word via an epenthetic glottal stop.

Primary stress falls on the initial syllable, secondary stresses land on subsequent odd-numbered syllables.

Plosives prenasalize after nasal vowels; nasals prestop after oral vowels. This introduces [p] and [ŋ] into the phonetic inventory.

When nasals are followed by /r/, an epenthetic homorganic plosive is inserted before the /r/. (Unresolved: what happens to /nr/ after oral vowels?)

All consonants labialize after rounded vowels.

All non-r consonants palatalize before front vowels. /t/ goes further, becoming [t͡ʃ] before any front vowel higher than /æ/.

/m t/ velarize before back vowels. /n/ becomes [ŋ] before back vowels. /k/ becomes [q] before any non-high back vowel.

/k/ and /t/ don't like to occur as onsets in adjacent syllables, so they morph. The plosive belonging to the more stressed syllable remains as is (subject to the above mentioned allophony) while in the unstressed syllable, /t/ -> [ʔ] and /k/ becomes [x], which in turn may labialize and/or palatalize|uvularize depending on the following vowel. (Hmm, /t/'s allophony here may be at odds with the epenthetic word-initial glottal stop rule--to resolve.)

Long oral vowels cause following plosives to become voiced.

[ol bofosh]

An experimental something:

/m m̥ n n̥/ <m hm n hn>
/p pʼ t tʼ k kʼ q qʼ/ <b p d t g k q c>
/p͡ɸ p͡ɸʼ t͡s t͡sʼ k͡x k͡xʼ/ <pp bb tt dd kk gg>
/l ɹ/ <l r>

/a e ɨ o ʉ/ <a e i o u>

CCVVCCC

Edit: long vowels
/aː ɛː iː ɔː uː/ <â ê î ô û>

[cromulant]

Inventory of 21V5C consonant allophones.

Code: Select all

                                t   tʷ   tˠ   tˠʷ    t͡ʃ   t͡ʃʷ   c   cʷ   k   kʷ   q   qʷ  ʔ
                                d   dʷ   dˠ   dˠʷ    d͡ʒ   d͡ʒʷ   ɟ   ɟʷ   g   gʷ   ɢ   ɢʷ
mb                             nd  ndʷ  ndˠ  ndˠʷ   nd͡ʒ  nd͡ʒʷ  ɲɟ  ɲɟʷ  ŋg  ŋgʷ  ɴɢ  ɴɢʷ
                                                                ç   çʷ   x   xʷ   χ   χʷ
                                                                ʝ   ʝʷ   ɣ   ɣʷ   ʁ   ʁʷ
 m   mʷ   mʲ   mʲʷ   mˠ   mˠʷ   n   nʷ                          ɲ   ɲʷ   ŋ   ŋʷ
pm  pmʷ  pmʲ  pmʲʷ  pmˠ  pmˠʷ  tn  tnʷ                         cɲ  cɲʷ  kŋ  kŋʷ
bm  bmʷ  bmʲ  bmʲʷ  bmˠ  bmˠʷ  dn  dnʷ                         ɟɲ  ɟɲʷ  gŋ  gŋʷ
                                ɾ   ɾʷ
                                r   rʷ
                                l   lʷ

[Chengjiang]

Tazurin:

/m n ɲ ŋ/
(/mb nd ɲɟ ŋɡ/)*
/p t tʃ k/
/f s/ (/ʃ/)*
/β z ʝ ɦ/
/l/

/ɪ ɐ ʊ iː eː aː oː uː/

|N|

*[mb nd ɲɟ ŋɡ] normally occur as the realizations of /Nβ Nl Nʝ Nɦ/, but they occur word-initially in loanwords and in a restricted set of native words as the result of consonant mutation. [ʃ] occurs in native words as an allophone of /s/ before palatal/palatoalveolar consonants and for some speakers before /ɪ iː/, but also occurs in other environments in loanwords.

/l/ > [ɾ] intervocalically, > [lʲ] before palatal/palatoalveolar consonants
|N| > nasal with same POA as following consonant; realized as [ŋ] word-finally before a pause or a vowel-initial word

Syllables are of the form (C)V(s/l/N); syllables may lack an onset only word-initially.

The romanization is to some extent subphonemic, distinguishing certain allophones:

<m n ny g>
<mb nd nj ng>
<p t ch k>
<f s> (<sh> in loanwords only)
<v z j h>
<l> (<r> between vowels)

<i a u ī e ā o ū>

<m> before labial consonants, <n> otherwise except in the sequence /Nŋ/ [ŋː], which is written <gg> to distinguish it from [ŋɡ] <ng>

Example: <tazurin> = /tɐzʊlɪN/ = [tɐzʊɾɪŋ]

[Nortaneous]

Mesotibetan:

/pʰ p ʰp b mb tʰ t ʰt d nd tʃʰ tʃ ʰtʃ dʒ ndʒ kʰ k ʰk g ŋg qʰ q ʰq ɴɢ/
/sʰ s ʰs z ʃʰ ʃ ʰʃ ʒ xʰ x ʰx ɣ χ ʁ h/
/n n̥ ɲ ɲ̥ ŋ ŋ̥/
/r j w/
/a e ɤ o i ɯ u/
/˥ ˧ ˎ ˧˥ ˩˧ ˏ˧/, word-level

Syllable structure: {(F|L|wF)C|(w)CF}V{(F|L|wF)C|(w)CF}, word-final clusters are more likely in monosyllables
Word structure: (n)S(S) + (S) affix

/nw/ [m], /w/ fricates initially before a consonant. Aspirated/unaspirated contrast is neutralized in coda. nF > nPF.

ntʃʰɯwɣa˥ sʰɤ˧ xedʁ˧˥ ŋgan̥iʰʃtˏ˧ waχˎ ʒmbɯ˧ ɸʃkɤɲrɤ˩˧ rɯ˩˧ βdʒa˧

[CatDoom]

Here's something I've been playing with for the next phase of the collaborative conworlding project. Let's call it Pre-Classic Ngmu' for now.

Consonants:

/m n n͡m ɳ ŋ ŋ͡m/
/b~β t t͡p ʈ k k͡p ʔ/
/ʂ h/
/ɻ ɰ ɰᵝ/

ɰᵝ is a compressed labio-velar approximant similar to the Japanese /w/

Vowels:

Ngmu' has a vertical vowel system; it has four vowels specified only for height, with backness and rounding being conditioned by the preceding consonant. Retroflex and velar consonants (with the exception of /ɻ/) cause backing, while alveolar consonants cause fronting. Labial consonants condition compressed lip rounding in non-open vowels. Co-articulated labial-alveolar and labial-velar consonants (including ɰᵝ) cause rounding in addition to conditioning vowel backness.

/ɨ/ ~[y]~[ɨ]~[ʉ]~[ɯ]~
/ɘ/ [e]~[ø]~[ɘ]~[ɵ]~[ɤ]~[o]
/ɜ/ [ɛ]~[œ]~[ɜ]~[ɞ]~[ʌ]~[ɔ]
/a/ [æ]~[a]~[ɑ]

Each vowel also has contrasting oral and nasal forms.

Suprasegmentals:

Ngmu' has a mixed tone system distinguishing 4 level tones and 7 contour tones encompassing a total of 5 registers. The registers are numbered according to the system used for the Oto-Manguean languages, where 1 represents the highest pitch and 5 the lowest.

Level tones: 22 33 44 55
Rising tones: 21 32 43
Falling tones: 12 23 34 45

Each tone also combines freely with contrasting plain, creaky-voiced, breathy-voiced, and strident phonation.

Syllable Structure:

Ngmu' syllables are uniformly of the form CV.

[cromulant]

CatDoom, this is the second time in a week you've given me a not-so-serious idea.

Anti-Kuty

/t k͡t k/
/tʼ kʼ/
/ⁿt ⁿk/
/tⁿ kⁿ/
/m n͡m n ŋ͡n ŋ n͡m/
/r/

/i ɨ u ə o a i: u: w͡a w͡i a͡w i͡w/

V(r)C

Tones:

Low: 1
Low rising: 12
Low falling: 21
Mid: 3
Mid rising: 34
Mid falling: 43
High: 4 (allotone of Mid Rising)

[Znex]

My updated phonology for Proto-Oeg:

/m n̪ n ŋ/ <m nh n g>
/p~p' t̪~t̪' t~t' k~k' ʔ/<p/f th t/s k/x c>
/l j w/ <l j v>

/i y u/ <i u w>
/e ø o/ <e eo o>
/ə/ <y>
/ɛ œ ɔ/ <ae oe ao>
/a ɑ/ <a aa>

/øy ou/ <eu ow>
/ai ei au/ <aj ej aw>

Word structure: (C)(C)V(T)(C) or (C)V(T)(C)(C)

Tones: 5 level tones; toneless words take the previous word's tone

[Cúlro]

(We seem to be having a run of tonal languages at the moment, so..)

Had an idea for a phonology while working on my main project. I like tonal languages, but they are horrible to romanise - I'm not a fan of nonintuitive romanisations (like 'Hmoob' style tone consonants, or acute accents for level tones); I like Mandarin pinyin style tone diacritics, where the diacritic looks like the tone contour. So, I decided to make a phonology based on that principle - tones are distinguished by contour shape only and can be represented with diacritics easily in the orthography. Here it is:

Vą̈i-rŵng-tẏ̨ Phonology

Consonants
Voiceless stops: /p t k/ p t k
Prenasalised stops: /^mb ⁿd ^ŋg/ ^mb ⁿd ⁿg
Voiceless affricate: /ʧ/ c
Prenasalised affricate: /ⁿʤ/ ⁿj
Voiceless fricatives: /f s/ f s
Voiced fricatives: /v ð/ v d
Liquids: /l r/ l r
Nasals: /m n ŋ/ m n ng

Vowels
/i u e o a ɤ ɯ/ i u e o a y w
All vowels can be contrastively nasalised (indicated with an ogonek ą), and preaspirated (indicated by a preceding h), or both.

Tones
There are seven tones, which can appear on all vowels, and combine freely with both nasalisation and preaspiration. The following description uses a 5 level transcription, where 1 = lowest, 5 = highest:
Level: ā 44
Rising: á 35
Falling: à 51
Dipping: ǎ 214
Peaking: â 352
Broken: ä 11 - with concomitant glottalisation, realised as a glottal catch~creaky voice
Checked: ȧ 3 - shorter than the other tones, terminated by glottal stop
Toneless: a - the equivalent of an unstressed syllable, occurs in grammatical particles and clitics, never on diphthongs.

Phonotactics
Syllable structure: (C)(i,u,r)V(i,u,e,o)(S), where V can be any vowel, S ={f s r m ng} or a glottal stop in checked tone syllables. Syllables with checked tone cannot have a coda consonant; the other six tones can co-occur freely with both open and closed syllables.
Tone and nasalisation are transcribed on the main vowel V.
Nasalisation is non-contrastive before nasal codas m ng and is not marked; vowels before nasal codas are allophonically nasalised (generally to a lesser degree than actual phonemic nasal vowels).
Preaspirated vowels can occur after p t k c f s l r m n ng and the glides i u.
Sequences of identical vowels cannot occur.
Prevocalic i u are realised as glides /j w/.
The only permitted onset clusters are p t k ^mb ⁿd ⁿg m n ng followed by r.
The glide u cannot follow f v.
Word-initial onsetless syllables are pronounced with an epenthetic glottal stop.
Back unrounded vowels y w cannot be preceded by the glide /w/ u, or followed by u o.
Vowel length is not contrastive; dipping, peaking and rising tones tend to be longer than level and broken tones, falling tone shorter than those, and checked tone shorter again; but apart from the checked tone, the difference are allophonic - by-products of the articulation of the tone rather than essential features of the realisation.
Diphthongs V(i,u,e,o) must be closing - ie vowels must appear in the order a > e o y > i u w and not both be of the same height group.
Thus the occurring diphthongs are ae ao ai au ei eu oi ou yi. In all dipthongs, the first element is longer and more prominent than the second.

Morphemes are usually monosyllabic, with compounding and affixation producing polysyllabic words.

There are no restrictions on consonant clusters at morpheme boundaries.

Prosody
There is light stress on the final syllable (unless the final syllable is a toneless clitic/suffix), realised by a slight increase in intensity, volume and a slight upstep in the absolute pitch of the 5 tone levels.

Allophony
In rapid speech, the complex tone contours (dipping and peaking) are often reduced. Most of the truncation occurs at the beginning of the contour, with the peak/trough value and the end preserved - ie the dipping contour can be realised as low rising 14, and the peaking contour as high falling 52. The extent of this reduction varies from speaker to speaker. The peaking tone usually undergoes less reduction, due to similarity of a reduced peaking tone 52 and the falling tone 51.
When a preaspirated vowel follows a voiced segment (r l m n ng u i) the voiced segment often undergoes partial devoicing, developing a voiced-to-voiceless contour.
The alveolars t ⁿd n are realised as dental when followed by the alveolar trill r.
The fricative s varies freely between [ɕ], [s] and [ʃ] before i, and [ʧ] and /ⁿʤ/ are in free variation with [ʨ] and /ⁿʥ/ respectively in the same environment.
Before i, l n are usually palatalised to [ʎ ɲ].
Nasalised /e o i u/ are usually realised as [ɛ̃ ɔ̃ ɪ̃ ʊ̃].

Preaspiration
Prevocalic h is analysed as preaspiration of the nucleus rather than postaspiration of the onset for three main synchronic reasons.
1. Infixes which appear after the onset appear between the onset and the h: pǫ̀ > pārǫ̀, but phǫ̀ > pārhǫ̀, not *phārǫ̀.
2. Words with apparent initial h are realised with the epenthetic glottal stop characteristic of onsetless words: hw̄r [ʔhɯr]
3. Glides u i /w j/ surface before the voiceless segment h; when the glides follow a voiced consonant this often results in an allophonic voicing contour during the glide from voiced to voiceless.
This prevocal realisation is presumably due to competition between a breathy-voiced 'spread glottis' feature on the vowel, and the necessity of modal voice to properly articulate the tonal contour, resulting in a sequence of voicelessness followed by modal voice.

EDIT: got the diphthong list wrong.

[CatDoom]

He he, "Hmoob." It's always a disappointment when the language turns out to not be nearly as hilarious as the orthography makes it look.

[cromulant]

While we're on the topic of tone romanization, here's the Kuty way:

High (4): á
High Rising (34): a̋
High Falling (43): â
Low (2): à
Low Rising (12): ǎ
Low Falling (21): ȁ

Any objections?

[Cúlro]

While we're on the topic of tone romanization, here's the Kuty way:

High (4): á
High Rising (34): a̋
High Falling (43): â
Low (2): à
Low Rising (12): ǎ
Low Falling (21): ȁ

Any objections?

Your rising and falling tones - wouldn't it be more intuitive to use â ǎ for the two falling tones and a̋ ȁ for the two rising tones? Then diacritic shape (v vs //) would distinguish contour (rise vs fall) and diacritic direction would distinguish pitch. It just seems counterintuitive to me to have the low rising diacritic be a flipped over version of the high falling instead of the high rising.
I think I see your reasoning though - each diacritic has a / or \ at its left side to distinguish if it's a high or low tone, and then the right hand side / or \ or nothing symbolises the contour shape. That's quite systematical and logical and a good way of doing it - the diacritic shape even mirrors the tone names. It just struck me that having v or ^ for falling and // or \\ for rising tones, with the direction they point distinguishing high or low would be more intuitive. But I guess what's intuitive is different for different people and both work fine.

[Cúlro]

Btw, anyone see anything wrong with the phonology I posted three posts back?

If there's any glaring problems I'd like to post them now before I do too much other work on it.

Thanks!

[CatDoom]

Giving it a quick glance, about the only issue I have is that I'm not sure about the term "preaspirated vowel." Jalapa Mazatec has vowels that are typically realized as voiceless for about half of their length and then modally voiced for the rest, which seems to be more or less what you're getting at, but the preferred terminology for this phenomenon still seems to be "breathy-voiced."

[Cúlro]

True, it's phonemically partial breathy voice, but I wanted a term that reflects the phonetic realisation more obviously. It's not a standard term, no, but at least it's shorter than "breathy voice realised on the first part of the vowel." It's the same reason I don't like purely phonemic transcriptions of conlangs (unless there's very little allophony) - it requires too much remembering of phonological rules on the part of an incidental reader, who is not really invested enough to put enough effort in to remembering rules that they are not going to need again once they've finished reading the post. If I just referred to it/notated it as breathy voiced with a note at the beginning on the exact realisation, would casual readers (ie everyone who might realise it) remember it as they read transcriptions? I might use it in the grammar as a language specific term, but obviously I'll use breathy voice in the description of the phonological description.

Thanks!

[Cúlro]

Forgot to ask this, but what does Vą̈i-rŵng-tẏ̨ (the language, not the word) remind/make you think of? Where would you imagine it spoken?

[CatDoom]

Hmm... between the complex tone and phonation contrasts, the unrounded back vowels, and the prenasalized stops, I think it wouldn't be too out of place somewhere in southeast Asia. The voiced fricatives might stand out, but some of the languages in that area have pretty wild consonant inventories, so it might still fit in.

[Cúlro]

My updated phonology for Proto-Oeg:
....
Tones: 5 level tones; toneless words take the previous word's tone

That strikes me as a little unrealistic (and possibly impossible). AFAICR, no natlangs distinguish this many level tones. Many natlangs have a 5 level system, but that means that the various contour tones and level tones that exist all either begin or end on one of 5 levels eg Mandarin's 55, 35, 214, 51. I'm fairly sure that an inventory of 55, 44, 33, 22, 11 is both unattested and extremely unlikely. It's hard to see how it could develop, and it seems like the tones would not be distinctive enough - the pitch differences are very small, and I'd be surprised if the contrast of a 11 tone followed by a 55 tone and a 11 tone followed by a 44 tone would be reliably perceived by listeners.

I would expect/recommend the 44 and 22 tones to change to contours - 55 33 11 level tones and some contours is a well attested system.

[----]

It's not *really* that unbelievable. Several languages distinguish 4 level tones, and at least one source describes a Tai-Kadai language by the name of Gaoba Dong with 5 level tones. Trique apparently has 5 level tones as well in some varieties.

the pitch differences are very small, and I'd be surprised if the contrast of a 11 tone followed by a 55 tone and a 11 tone followed by a 44 tone would be reliably perceived by listeners.

Tone numbers do not describe absolute pitches. They describe pitches in the context of each other.

[CatDoom]

Usila Chinantec also has 5 level tones; it's rare, but not unheard of. I suspect that context clues help distinguish near-homophones with similar tones.

[Nortaneous]

Black Miao, Dan, and Chicahuaxtla Triqui have five phonemic level tones. Ngamambo has five phonetic level tones, but only two phonemic tones. Chori has six phonetic level tones, but only three phonemic tones.

[Cúlro]

the pitch differences are very small, and I'd be surprised if the contrast of a 11 tone followed by a 55 tone and a 11 tone followed by a 44 tone would be reliably perceived by listeners.

Tone numbers do not describe absolute pitches. They describe pitches in the context of each other.

I know that and that was what I was talking about - pitch differences. I was saying that a contrast between two bisyllabic words, say, ta11m44 and ta11ma55 would be perhaps difficult to hear in isolation - ie if one of the words was spoken, would the hearer be able tell if it was ta11m44 or ta11ma55? In Gaoba Dong, each tone is realised shorter than the one below it, and the level tones are actually realised with slight contours - they end at equally spaced levels, but 11 22 and 55 all start more bunched together towards the mid level, and 22 even has a dip in the middle:

image.jpg (95.44 KiB) Viewed 4890 times

Also, I perhaps worded it poorly, but I meant no natlang distinguishes *just* this many level tones - no contours. Gaoba Dong has 3 rising contours and a falling contour as well as those level tones, and Usila Chinantec has contours too.

5 level tones are perceptually very close - I would expect them to spread out into contours if there were no contours already to maintain a contrast with. Are there any attested cases of 5 level tones as the only tones?