Post your conlang's phonology

[8Deer]

That works too. Your neutralisation of z/ɹ intervocalically is still a good change, for example.

Actually, isn't that what happened in Old or Proto Norse?

[Whimemsz]

That works too. Your neutralisation of z/ɹ intervocalically is still a good change, for example.

Actually, isn't that what happened in Old or Proto Norse?

It's happened a few times. Northern and Western Germanic is one instance. The other well-known one (actually I've never seen it described elsewhere, maybe it's just these two?) is Latin, which had *s > *z > r between vowels (thus genus~generis)

[Shihali]

My second conlang, Rimelsó:

Consonants:

m n ɲ ŋ
p t tʃ k q '
p^h t^h tʃ^h k^h q^h
p' t' tʃ' k' q'
^mb ⁿd ^ɲdʒ ^ŋg
(pf) (ts) (kx) (qχ)
(ts^h)
s ʃ h
(v) (z) (ʒ) (ɣ)
w y
r
(ɾ)
l ʎ

Vowels:
i(ĩ) u(ũ)
e o
ɛ(ɛ̃) ɔ(ɔ̃)
æ ɑ

Sounds in parentheses are allophones only.

Syllable structure:
(C)V(:)(N)(r|L)(C)
On the surface, Standard Rimelsó only permits r, l, or ľ plus a consonant in the coda; any nasal element other than a single syllable-final nasal nasalizes the vowel and falls. The semivowels w and y are not permitted at the end of a syllable; historically they formed diphthongs that became monophthongs, but they resurface if a suffix beginning with a vowel is added.
Vowels are also nasalized if they precede a prenasalized consonant.

The Tési dialect has merged the prenasalized series into the tenuis series, but retains the nasalized vowels; nasalization is phonemic in Tési.

This phonological inventory doesn't cover the allophonic lenition within words; I'm still working out the rules governing clusters, especially word-final clusters which have to shed consonants.

I have problems with the orthography. Not the sounds, which are easy enough to write although leading to horribly unintuitive results, like <boquuʼ> /mbɔhuːʔ/, but whether to write words more phonemically or morpho-phonemically. "To a man" is *ras-m, and pronounced /rɔ̃s/ in the standard dialect, but there are about equally good cases for spelling it <rǫs> (how it is pronounced), <rans> (how it is pronounced in conservative dialects), and <rasm> (unpronounceable in any dialect but clearly showing the morphology). I'm currently leaning towards a heavily morphophonemic orthography because a sensible phonemic orthography is still of little help to someone who doesn't know Rimelsó spelling conventions.

[cromulant]

An inventory:

Code: Select all

p        t        k
b        d        g   
    θ    s                  ħ
m   n̟   n
         ɾ
              j      w

i e a u o

Problematic? There are some unusual consonants (θ n̟ ħ) for such a small inventory, but I've got all the vanilla with no gaps.

EDIT: changed the IPA symbols for the interdental nasal.

[Wattmann]

An inventory:

Code: Select all

p        t        k
b        d        g   
    θ    s                  ħ
m   n̪   n
         ɾ
              j      w

i e a u o

Problematic? There are some unusual consonants (θ n̪ ħ) for such a small inventory, but I've got all the vanilla with no gaps.

Nothing is unusual with those consonants - inventories don't have needs based on size.
What do you mean by vanilla? I like strawberry more

[cromulant]

An inventory:

Code: Select all

p        t        k
b        d        g   
    θ    s                  ħ
m   n̪   n
         ɾ
              j      w

i e a u o

Problematic? There are some unusual consonants (θ n̪ ħ) for such a small inventory, but I've got all the vanilla with no gaps.

Nothing is unusual with those consonants - inventories don't have needs based on size.

That's not quite true. Less common consonants tend to be found in larger inventories, and absent from smaller ones.

http://wals.info/chapter/19

^The last two paragraphs are the relevant part.

What do you mean by vanilla? I like strawberry more

By that I mean the plain, basic phonemes like /p t k m n s/ without which I probably couldn't have the more exotic stuff. For example, I probably wouldn't have /n̪/ without /n/, or /θ ħ/ without /s/. Languages overwhelmingly fill up all the vanilla before moving on to the strawberry.

[Pole, the]

By that I mean the plain, basic phonemes like /p t k m n s/ without which I probably couldn't have the more exotic stuff. For example, I probably wouldn't have /n̪/ without /n/

That's true only for phonemes.
E.g. Polish /n t d/ are usually [n̪ t̪ d̪] respectively (and may be described as /n̪ t̪ d̪/ without /n t d/).

[cromulant]

The inventory I posted has a phonemic alveolar/interdental contrast in the nasals and fricatives. Sorry, I /n̪/'ed when I should have /n̟/'ed.

[Tropylium]

Unnamed

/p t k kʷ/
/ts tɕ/
/s ð ɕ x ɣ xʷ/
/l ɹ j w/
/m n ɲ ŋ ŋʷ/

/a e i o u ɨ/
/ai au/
All vowels have a length contrast except /ɨ/

Allophony:
Stops before a stressed syllable are aspirated.
/ɣ/ is /ŋ/ intervocalically.
/*ɣʷ/ merged with /w/.
/ð/ is [ɹ] intervocalically.
/x/ is pronounced /h/ by many speakers intervocallically.

Phonotactics:
(C)(C)V(C), where the second consonant is a fricative or a resonant (haven't fully worked out which clusters are allowed).

This language is an isolate spoken in Siberia, so I'm trying to get a Siberian flavour with it.

the problem is your /ð/ - a /θ/ is far more likely there (it's not an actual problem, but somewhat bugging - do consult with other members)

Also, x~ɣ is the only voicing contrast in your language – I'd be a bit hesitant about that as well, although there's nothing wrong with it per se.

Having /ð/ without /θ/ and /x ɣ/ as the only voicing contrast are both attested in languages of western Siberia. The consonant inventory of Northern Mansi, for example, is:
/p t k kʷ/
/s ʃ ɕ x xʷ/
/m n ɲ ŋ ŋʷ/
/l ʎ r j ɣ w/
Not that there's anything wrong with having /z/, either (but I think it's quite a bit rarer than /ɣ/ in Siberia; it's found in Mongolic at least, but eg. Northern Turkic languages tend to have *z → s.).

OTOH what is not all that Siberian in feel is a (C)(C)V(C) syllable structure, and phonemic difthongs. I think pretty much all languages around there rather have (C)V(C)(C) and monofthongs only.

…Also if both /w/ and /ŋ/ are independant consonants tho, how is /ɣʷ/ "merged with" the former, but /ɣ/ is "allophonically" the latter?

[Wattmann]

…Also if both /w/ and /ŋ/ are independant consonants tho, how is /ɣʷ/ "merged with" the former, but /ɣ/ is "allophonically" the latter?

Simple?
*ɣʷ w
ɣ ~ ŋ

[Wattmann]

http://wals.info/chapter/19

^The last two paragraphs are the relevant part.

What do you mean by vanilla? I like strawberry more

By that I mean the plain, basic phonemes like /p t k m n s/ without which I probably couldn't have the more exotic stuff. For example, I probably wouldn't have /n̪/ without /n/, or /θ ħ/ without /s/. Languages overwhelmingly fill up all the vanilla before moving on to the strawberry.

This one corrected stands.

I tend to leave various things out of the vanilla before going strawberry :3

[Tropylium]

Languages overwhelmingly fill up all the vanilla before moving on to the strawberry.

I tend to leave various things out of the vanilla before going strawberry :3

Well, cromulant's statement is probably best applied to various MOAs separately; it's nothing too surprizing to have /p t ts tʃ k kʷ/ × a few phonations, and yet not more than 3-4 sonorants (out of the "vanilla" /m n l r j w/).

It is unusual (altho not necessarily completely impossible) to have /kp/ without /p/, /ŋ/ without /m/, /ɭ/ without /l/, etc.

[Wattmann]

Languages overwhelmingly fill up all the vanilla before moving on to the strawberry.

I tend to leave various things out of the vanilla before going strawberry :3

Well, cromulant's statement is probably best applied to various MOAs separately; it's nothing too surprizing to have /p t ts tʃ k kʷ/ × a few phonations, and yet not more than 3-4 sonorants (out of the "vanilla" /m n l r j w/).

It is unusual (altho not necessarily completely impossible) to have /kp/ without /p/, /ŋ/ without /m/, /ɭ/ without /l/, etc.

I tend to leave out labials, and make nasals either only barely phonemic or batshit rich (five POAs, glottalisation, length)

[Shrdlu]

Made a small phonology based on the stuff written above and left out /n/ to give it a certain flavour.

Code: Select all

/p[p] pv[p_>] pp[p:] t[t'] tt[t:] ts[ts] c[tS] cc[tS:] k[k] kv[k_>] kk[k:] kw[k_w]/
/m[m] g[n'] gw[n'_w] l[l] r[r] j[j] w[w]/
/s[s] ss[s:] h[h] hw[h_w]/

(C)V(C)(C)

[finlay]

…Also if both /w/ and /ŋ/ are independant consonants tho, how is /ɣʷ/ "merged with" the former, but /ɣ/ is "allophonically" the latter?

The way I understand it is, ɣʷ doesn't exist anymore because it merged with w, while ɣ and ŋ are still distinctive in the initial and/or final position. The proper terminology for the second type is a "neutralization", by the way, if two sounds are distinctive in one position and not distinctive in another. The best RL example of this is when you have a language with /n m/ as its nasal phonemes and [ŋ] only before velars. It's theoretically impossible to assign [ŋ] as /n/ or as /m/, because neither [n] nor [m] appears before a velar – the two phonemes are neutralized.

If you have two sounds, ɣ and ŋ for instance, one of which appears in one position and the other in another position, then you have a very basic example of one phoneme with two allophones. As it stands, I know that ɣ and ŋ aren't distinctive intervocalically for 8Deer's language, but I don't know about other positions – ŋ is a very commonly disallowed in initial position, for example, and it could be that ɣ is disallowed in the final position, making it one phoneme with [ɣ] initially and [ŋ] interv./finally. You could even have the distinction only crop up in one place, such as the final position.

[Tropylium]

…Also if both /w/ and /ŋ/ are independant consonants tho, how is /ɣʷ/ "merged with" the former, but /ɣ/ is "allophonically" the latter?

The way I understand it is, ɣʷ doesn't exist anymore because it merged with w, while ɣ and ŋ are still distinctive in the initial and/or final position. The proper terminology for the second type is a "neutralization", by the way, if two sounds are distinctive in one position and not distinctive in another. The best RL example of this is when you have a language with /n m/ as its nasal phonemes and [ŋ] only before velars. It's theoretically impossible to assign [ŋ] as /n/ or as /m/, because neither [n] nor [m] appears before a velar – the two phonemes are neutralized.

If you have two sounds, ɣ and ŋ for instance, one of which appears in one position and the other in another position, then you have a very basic example of one phoneme with two allophones. As it stands, I know that ɣ and ŋ aren't distinctive intervocalically for 8Deer's language, but I don't know about other positions – ŋ is a very commonly disallowed in initial position, for example, and it could be that ɣ is disallowed in the final position, making it one phoneme with [ɣ] initially and [ŋ] interv./finally. You could even have the distinction only crop up in one place, such as the final position.

Altho if they do contrast, I see no reason to analyze medial [ŋ] as a neutralization of /ɣ ŋ/ rather than as simply /ŋ/; what association does this have with /ɣ/ anymore? Couldn't we equally well claim that it's actually medial [x] that is a neutralization of /ɣ x/?

I could understand if the point was a change from medial *ɣ to /ŋ/, but this change was explicitly listed as allophonic and not historical (unlike the *ɣʷ → w change).

[finlay]

…Also if both /w/ and /ŋ/ are independant consonants tho, how is /ɣʷ/ "merged with" the former, but /ɣ/ is "allophonically" the latter?

The way I understand it is, ɣʷ doesn't exist anymore because it merged with w, while ɣ and ŋ are still distinctive in the initial and/or final position. The proper terminology for the second type is a "neutralization", by the way, if two sounds are distinctive in one position and not distinctive in another. The best RL example of this is when you have a language with /n m/ as its nasal phonemes and [ŋ] only before velars. It's theoretically impossible to assign [ŋ] as /n/ or as /m/, because neither [n] nor [m] appears before a velar – the two phonemes are neutralized.

If you have two sounds, ɣ and ŋ for instance, one of which appears in one position and the other in another position, then you have a very basic example of one phoneme with two allophones. As it stands, I know that ɣ and ŋ aren't distinctive intervocalically for 8Deer's language, but I don't know about other positions – ŋ is a very commonly disallowed in initial position, for example, and it could be that ɣ is disallowed in the final position, making it one phoneme with [ɣ] initially and [ŋ] interv./finally. You could even have the distinction only crop up in one place, such as the final position.

Altho if they do contrast, I see no reason to analyze medial [ŋ] as a neutralization of /ɣ ŋ/ rather than as simply /ŋ/; what association does this have with /ɣ/ anymore? Couldn't we equally well claim that it's actually medial [x] that is a neutralization of /ɣ x/?

I could understand if the point was a change from medial *ɣ to /ŋ/, but this change was explicitly listed as allophonic and not historical (unlike the *ɣʷ → w change).

If they alternate, then you've definitely got a neutralization or allophony situation. That's pretty much how it works...

[sirdanilot]

An inventory:

Code: Select all

p        t        k
b        d        g   
    θ    s                  ħ
m   n̟   n
         ɾ
              j      w

i e a u o

Problematic? There are some unusual consonants (θ n̟ ħ) for such a small inventory, but I've got all the vanilla with no gaps.

EDIT: changed the IPA symbols for the interdental nasal.

The only remarkable things here are ħ and perhaps the absence of t̪ d̪ (thouɡh they could have become θ or merɡed with their alveolar counterparts or something). But how to explain the pharyngeal? Perhaps the pharyngeal used to be part of a series of pharyngealized stops, which have since disappeared (perhaps leaving a lot of a's and o's behind, as pharyngeals like to do) except for the ħ. This seems the most plausible explanation. One would perhaps expect /q/ and /ɢ/ then, or remnants of those.
Also, /f/ is missing but why the hell no if it floats your boat?

[Whimemsz]

But how to explain the pharyngeal? Perhaps the pharyngeal used to be part of a series of pharyngealized stops, which have since disappeared (perhaps leaving a lot of a's and o's behind, as pharyngeals like to do) except for the ħ. This seems the most plausible explanation. One would perhaps expect /q/ and /ɢ/ then, or remnants of those.
Also, /f/ is missing but why the hell no if it floats your boat?

I don't see how either of these is a problem. You can have a pharyngeal fricative without a pharyngealized stop series (the UPSID list of languages with voiceless pharyngeal fricatives includes a number with no other pharyngeal(ized) consonants, although in a lot of cases their analysis seems to be based on older studies and not actually correct...nonetheless, see for instance Kurdish)--what's odder is to have such a situation in a language with a relatively small phoneme inventory overall. But I still think it's perfectly doable. And lacking /f/ while having other fricatives isn't even a little bit odd. In fact, the fricative inventory here is close to that of Karajá, which also serves as a good illustration that smallish inventories can still be fucking nutso.

[Rui]

Just had a strike of inspiration for a phoneme inventory:

/m n ɳ/ <m n ṇ>
/pʰ tʰ tʼ ʈʰ kʰ kʼ ʔ/ <p t d ṭ k g ʔ>
/ʈʂʼ tɬ’ kx/ <ḍ l q>
/θ ð ʂ χ/ <s z ṣ x>
/j ɭ ɰ/ <y r w>

/ɑ e i o u/ <a e i o u>
/ɑ: i: u:/ <ā ē ō>

A few things I'm debating about:
-getting rid of /ð/. I feel like it doesn't fit in with this inventory. But that's just me. Also the whole "only one set of voicing contrasts" thing
-adding /f/, to better explain the lack of /pʼ/...I know /pʼ/ isn't terribly common, even in langs with ejectives, but I feel like it would be better to have /pʼ/ > (pf >) f. My only hesitation is that contrasting /f θ/ would be weird [yes, I'm aware standard English is an obvious example of a lang that does, but English also has like 58731935103 fricatives anyway...]

Also: what about /ʈʂʼ/? Wikipedia says it's attested in 2 languages, but then when I go to their pages, one doesn't have it listed in its phoneme inventory (it has /tʃʼ/), and one has a sparse page that doesn't even have a consonant inventory on it...

Also also: is it weird to have /ɳ/ without /ŋ/?

[Pole, the]

Also also: is it weird to have /ɳ/ without /ŋ/?

I don't think so.

Also, all other thing seems ok. /ð/ would fit better changed to /ð̞/, though.

Edit: /tʰ tʼ/ are dental here, aren't they?

[Whimemsz]

Also: what about /ʈʂʼ/? Wikipedia says it's attested in 2 languages, but then when I go to their pages, one doesn't have it listed in its phoneme inventory (it has /tʃʼ/), and one has a sparse page that doesn't even have a consonant inventory on it...

Wikipedia is stupid, and contradicts itself in several places on this question. It also conflicts, sort of, with UPSID's database. Anyway, look no further than the Mamean languages (and some other Mayan languages, such as Jacaltec) for several examples of languages with /ʈʂʼ/.

EDIT:

/ð/ would fit better changed to /ð̞/, though.

No it wouldn't. Unless you've got contrastive interdental and dental non-sibilant fricatives, there's no need to specify the contrast in phonemic notation. And in any case it's perfectly natural to have interdental fricatives corresponding to a dental stop series (or even, in addition to alveolar stop and fricative series, or some other such combination)

EDIT2:

Also also: is it weird to have /ɳ/ without /ŋ/?

It's unusual, but certainly not impossible or unattested (take Hungarian)

[Wattmann]

Also also: is it weird to have /ɳ/ without /ŋ/?

It's unusual, but certainly not impossible or unattested (take Hungarian)

Uh, NO?
That's palatal, this' retroflex.

[Whimemsz]

Shit, you're right. It's not my fault the two symbols are so similar, though, someone should do something about that >:|

Anyway, my actual point is still correct (see Pashto)

[Wattmann]

Shit, you're right. It's not my fault the two symbols are so similar, though, someone should do something about that >:|

Anyway, my actual point is still correct (see Pashto)