zompist bboard

I am currently working on a (possibly very unrealistic) conlang for an altenate earth that lacks nasals and labials. The only semivowel it has is /j~i/. I want to derive a more realistic daughter language from it that has a few labials and/or nasals. I was wondering if there are any attested sound changes that can derive labials or nasals from scratch, short of borrowing.

There’s a plenty of ways to derive labials and nasals. ☺

For example, rhinoglottophilia can produce nasals out of guttural consonants and labials can arise from velars before rounded vowels.

Oh, yeah. It doesn't have any rounded vowels either...

You could always employ dissimilation: tata > tapa, kaka > kapa. Additionally voiced stops could become nasals conditionally or unconditionally.

There's usually either a rounded /o/ or a rounded /u/. I can't think of any languages off the top of my head that have no rounded vowels whatsoever. Not even Japanese is without rounded vowels. Its u, which is often cited as unrounded, in fact is compressed rounded as opposed to protruded rounded.

vec wrote:There's usually either a rounded /o/ or a rounded /u/.

If you want to make this an absolute argument, change it into the phonetic level and talk about [o] or . There are languages which are analysed with minimalistic vowel systems lacking rounded vowel phonemes but I've never seen a language that would lack rounded vowel phones. Moreover, languages that can be seen as lacking rounded vowels always appear to have at least one rounded consonant phoneme. So for example Wichita is a language with no labials and seemingly no rounded vowels, but it has both labiovelar consonant phonemes /kʷ w/ and rounded vowel phones [o o:]

Yeah, good point.

Every language likely has a -like sound or a [o]-like sound or both. One of them can be unrounded or both, but not neither. If you only have one back vowel, it's probably gonna be rounded phonetically, at least in a lot of contexts. If you don't have phonemic back vowels, a la Ubykh, you sure as hell have rounded back vowels phonetically.

What I'd take from this, Clearsand, is that if you're happy to have your proto-lang be unrealistic, go for no rounded vowels and know that your back vowels will very likely round spontaneously and early once your sound change processes "start". Then derive your labial consonants from there.

for rounded vowels: do a Cheyenne-like vowel shift:

a > o
e > a (except adjacent to /j/ where it is preserved)
ej > e

so now you've got /o/. You could have it raise to next to /k/ as well.

What's the actual phoneme inventory as it stands right now? We could probably help you more with a full view.

Zju wrote:You could always employ dissimilation: tata > tapa, kaka > kapa.

I don't think that happens… /p/ is not quite "the opposite" of either /t/ or /k/, which is what dissimilation usually works with. I'd actually expect *tata > *taka or *kaka > *kata to be more likely (this is at least attested in one of the Oceanic languages IIRC).

In the languages that mostly lack labial consonants historically (Na-Dene, Iroquioan), loaning and labiovelar fronting have been the main pathways of introducing labials. E.g. in varieties of Athabascan, *w > *p.

Another route is interdental fronting: e.g. *θ > /f/, *ðr > *vr > /br/ in Latin, *ts > *tθ > /p/ in Mountain Slavey.

Zju wrote:Additionally voiced stops could become nasals conditionally or unconditionally.

I am skeptical on if this would occur in a language that lacks nasal consonants, either.

I can't think of any languages off the top of my head that have no rounded vowels whatsoever.

Here's two: Matsés (/i ɨ ɯ e ɤ a/, explicitly claimed to hold on the phonetic level), Alekano (/i ɯ e ɤ a/).

"There must be a rounded vowel" is only a statistical universal; since both /ɯ/ without /u/ and /ɤ/ without /o/ are possible, it's also clearly possible that there might be a language that has both of these rarities.

Tropylium wrote:Here's two: Matsés (/i ɨ ɯ e ɤ a/, explicitly claimed to hold on the phonetic level), Alekano (/i ɯ e ɤ a/).

Oh yes, those seem to be languages that have experienced some universal loss of rounding or something comparable to that. It's just a shame that there's nothing definitive around on the exact nature of the back vowels in these languages. The Wikipedia articles are very cursory on the subjects as is the SIL phonology sheet for Alekano. Even the phonology section of Fleck's Matsés grammar has some puzzling statements in place of a proper analysis such as (p. 91, emphasis mine):

The /o/ ([ɤ]) and /u/ ([ɯ]) in Matses sound just a bit different than in Spanish or English, differing at least in that the Matses vowels are unrounded.

It's only a qualitative assessment, but to me [o] and [ɤ] and likewise and [ɯ] sound more than just a bit different with each other. Something's strange here and I'm not sure if it's only the writing practice or also something in the analysis. Based on a statement like that the back vowels could equally as well be only partially unrounded.

Clearsand wrote:I am currently working on a (possibly very unrealistic) conlang for an altenate earth that lacks nasals and labials. The only semivowel it has is /j~i/. I want to derive a more realistic daughter language from it that has a few labials and/or nasals. I was wondering if there are any attested sound changes that can derive labials or nasals from scratch, short of borrowing.

/ki/ → /kɨ/ → /kɯ/ → /ku/ → /kuə/ → /kwə/ → /kʷə/ → /pə/

I think I will just have a vowel shift for my bizarre vowel phonology and round the back vowels. /i, ɪ, e, ɛ/ > /i, u, ɔ, a/ (Plus a few more conditional changes for additional vowels) The labials will just stem from /u/ and from the /f/. I'm still not sure what to do about nasals, though. In what conditions would another phoneme spontaneously nasalize?

Clearsand wrote:/f/

I thought you said you didn't have any labials.

Yeah... I didn't exactly have the phonology at hand when I wrote the first post and I had forgotten that I had the labiodental fricative because it was such a rare sound in the language. Besides, it's a Labiodental. I'm sure it doesn't have nasals, though. Sorry for the confusion.

But it's still labial.

what's the inventory?

Wow...A language without these is really curious to me. I've imagined for many times how would one sound/feel when speaking one.
But just to be a little uncreative...my source is of course incomplete.

*(t)ʂ(ʰ)u > (t)ʂ(ʰ)wu > (t)ʂ(ʰ)ʋu > (p)f(ʰ)u (Still effective when /u/ is a semivowel, in which case it's deleted. Notably in Jin Chinese; first step also seen in several other parts of China. Series of /ʃ/, not /ʂ/, may also undergo such changes.)
*hu > hwu > ɸu/fu; ɦu> ɦu > βu/vu (Still effective when /u/ is a semivowel, in which case it's deleted. Well-known and symbolic for Southern Chinese languages/dialects in common belief; indeed common there)
*kʰu >> hwu > fu (Still effective when /u/ is a semivowel, in which case it's deleted. /kʰ/ > /h/ is an independent sound change and isn't completely carried out. Guangzhou Cantonese)
*(ɣ/ɰ)u > wu > ʋu > vu (Still effective when /u/ is a semivowel, in which case it's deleted. Mandarin.)
*-u > -ʋ̩ (in many parts of China, especially when following /f/ or /v/. Independent sound change, and therefore can happen after the first four changes listed above)
*kʷ > p (Ancient Greek)
*dw > b (to Classical Latin)
*kw > kv (German)
*-i̯u̯i/-i̯u̯e/-i̯u̯əi > -u̯i (Mandarin, historically; not fully carried out. /j/ is deleted)
*-i̯uŋ > -im (Minnan, relic)
*-i̯uk > -ip (Minnan, relic)
*CM(l)ɯː/Cu:l > Cwʌi̯ (M for any labial including /w/ and /ʷ/, and C for any or no consonant. Old to Middle Chinese. OC *-l > Late OC *-i)
*-ɑ > -o > -wa (Minnan, as believed by some linguists), -ɑi > -oi > -wai (Some dialects of Cantonese and Hakka)
*/-ə/[-ɤ] > [-ɰʌ]
*o > wo (Japanese and Italian, historically)
*o > we (Spanish, historically)
*i̯o > y̯o (Mandarin, historically)
*ɯ > u (Minnan, some dialects)
*koː > kowo (Modern Japanese, mainly in singing )
*V.a > V.wa (V for any vowel. Japanese, in some speech)
*-Ok > -Ou̯ (O for any of /ɑ/ /u/. Cases of /-uk/ become modern [ou̯] but are relics. There was no /-ok/. Southern Mandarin, historically)
*Eu̯ > Ew > Ev/Ef (E for any of /a/ /e/ /i/, for no /u/ existed and /ou̯/ > /o:/ > /u:/. Greek, historically. /f/ when not immediately followed by a voiced sound in the word)
*-l > -w (Many...including some English dialects)
*-UK > -UKP (U stands for any rounded vowel and K stands for any velar consonant. Vietnamese)
*r > w (Some English dialects)


*-p, -t, -k > -m, -n, -ŋ (Old Chinese/Old Tibetic; sporadic but enough to set a group. Voicedness is not distinguished in final positions. e.g. no contrast between -p and -b)
*l-/r- > n- (seen in Southwestern Mandarin and Korean)
*(ɣ/ɰ)- > ŋ- (Mandarin, of Jiaoliao, Northeastern, etc.)
*b, d > mb, nd; g > ŋg > ŋ (Old to Late Middle Japanese; /g/ became fully nasalized because no /ŋ/ existed. [ŋ] remains today almost throughout the country, while prenasalization can still be seen in certain dialects)
also check en.wikipedia.org/wiki/Korean_phonology#Consonant_assimilation (like hak.ra > haŋ.na)
and, in PIE there's a nasal infix that can be inserted in some verb roots.
In Attic Greek, after certain verb person endings that end in vowel, an /n/ is inserted when what immediately follows is either a pause, nothing or a word beginning with a vowel or /h/, either to avoid consecutive vowels or perhaps to stop more gracefully...
In some dialects of Jin, some historical oral vowels can be nasalized without any condition (e.g. *-o > -õ). I couldn't really figure why..



Some reverse phonetic changes also come to my mind:

Middle Chinese *pʰim > Mandarin pʰin, MC *pim > Mandarin piŋ, MC *pi̯ap > Cantonese faːt, due to dissimilation (no syllable can both begin with and end in labial consonant. pi- > f under certain circms. is an irrelevant independent sound change)
PIE *p...kʷ > Latin kʷ...kʷ (sequence)
MC *m, *n, *ɲ, *ŋ > Minnan b/m, d/l/n, dz/l, g/ŋ (nasals are retained in nasalized syllables; d > l, and *ɲ > *ɲʑ > *dʑ > dz > (some dialects) l. /l/ is reported to be a lot like /d/)
MC *Pj- > Vietnamese T-, when the most prominent vowel in the nucleus is /i/ or /e/. (P and T are repectively for labials and alveolars. I mean loanwords here. Also seen in the Wenxi dialect of Jin Chinese)
Late OC *-um > MC *-i̯uŋ (dissimilation. Before OC short vowels, a /j/ glide is developed)

If a language has back vowels then generally it has rounded back vowels, and maybe in addition unrounded central or back vowels.

I suppose you could think of a language with only a height distinction and no backness distinction? Just high, middle and low. /i~ɨ~ɯ/ vs./e~o/ vs /a~ɑ/ or something, and then an additional feature to expend this threeway inventory (nasalization or something) But even then the inventory is probably going to 'spread out' to maximize the differences to /i/ /o/ /a/ I guess.

But something like /i/ /ə/ /a/ could be a kinda stable system I suppose and there's no real phonemic rounding there I guess.

sirdanilot wrote:But something like /i/ /ə/ /a/ could be a kinda stable system I suppose and there's no real phonemic rounding there I guess.

And then some of those vowels get backed by velars, and then rounded by the usual "wait, we should totes round our back vowels, guys!" thing.

Yea

Sorry guys but it's not for nothing that such a large proportion of the world's phoneme inventories are pretty standard and just variations on the same theme, a e i u o ptk nasals/liquids/semivowels and some contrasts thrown in over those and then perhaps one or two fun sounds to boot. It's just the easiest.