Whence aspirated stops and nasal vowels?

[Pabappa]

There arent a whole lot of languages that use pharyngealized vowels, but I imagine that pharungalized vowels could turn into nasal vowels unconditionally.

[Ketumak]

It's probably not impossible, but aspiration in consonant clusters is, as far as I understand it, a method of maintaining the distinction between the first and second sound, instead of leveling the distinction

It's perhaps worth noting that co-articulated consonants are different from clusters, both in terms of how they're articulated and how they sound.

These both make sense to me, thanks.

pharyngealized vowels

Phew, I never even knew there was such a thing! Interesting and simple to handle by the sounds.

[vokzhen]

pharyngealized vowels

Phew, I never even knew there was such a thing! Interesting and simple to handle by the sounds.

They're not common, even rarer than pharyngeal consonants. I they're present in a few Northeast Caucasian languages, Khoisan (where pharyngealized and epiglottalized vowels can contrast), as part of consonant-vowel harmony in some Northwest Coast languages, and in some Germanic (mostly or entirely from coda /r/). I'm not aware of them existing elsewhere, though I wouldn't doubt they do somewhere; UPSID lists Even (Tungusic), Hama (Omotic), and Neo-Aramaic as well.

[Travis B.]

pharyngealized vowels

Phew, I never even knew there was such a thing! Interesting and simple to handle by the sounds.

They're not common, even rarer than pharyngeal consonants. I they're present in a few Northeast Caucasian languages, Khoisan (where pharyngealized and epiglottalized vowels can contrast), as part of consonant-vowel harmony in some Northwest Coast languages, and in some Germanic (mostly or entirely from coda /r/). I'm not aware of them existing elsewhere, though I wouldn't doubt they do somewhere; UPSID lists Even (Tungusic), Hama (Omotic), and Neo-Aramaic as well.

Are they not also present in some Qiangic languages as well?

[Richard W]

Pharyngealisation of vowels (a.k.a retraction of vowels) seems to be fairly common in Tungusic languages, and has even spread to Mongolian, or at least the dialects most strongly associated with the name.

[HoskhMatriarch]

I hope I'm not too late to this thread to post here, but, about aspirated stops, if one way they can arise is through gemination, but most languages with aspirated stops and affricates don't have aspirated fricatives, what normally happens to the fricatives? Can they just turn into affricates?

[vokzhen]

They just stay geminates in most cases, I believe. /pp/ is [ppʰ] and /ss/ stays [ss], never gaining aspiration in the first place.

[Nortaneous]

Vowels with some sort of backing appear in some Qiangic languages, yes -- variably described as velarized, uvularized, or pharyngealized. I don't know if these are articulatorily different things. Japhug ejected velarization as a preceding consonant: jmŋo < *lmaˠŋ. (aŋ > o and lC > jC are regular. The consonant ejected was ɣ in some other cases.) Tangut is sometimes reconstructed with velarized vowels.

There's a wider part of Asia where +/-RTR distinctions exist on vowels, but where the distinctions can primarily be described in terms of quality -- for two examples, Yi and Mongolian. This may be true of some Qiangic languages, especially the ones that have been described as having vowel systems like /a ɑ e ə o i (y) u/ but frequent merging/alternation of the mid and high vowels -- other Qiangic languages have /a ə i u/ plus pharyngealization.

Search 'Hongyan Qiang' here.

[Tropylium]

about aspirated stops, if one way they can arise is through gemination, but most languages with aspirated stops and affricates don't have aspirated fricatives, what normally happens to the fricatives? Can they just turn into affricates?

They just stay geminates in most cases, I believe. /pp/ is [ppʰ] and /ss/ stays [ss], never gaining aspiration in the first place.

Another option is voicing. It's not too rare to find languages that contrast lack vs. presence of aspiration in stops, but voicing vs. voicelessness in fricatives. (Two good examples are Icelandic and Nivkh.) Often the unaspirated stops may allow voiced allophones in some positions — but they do not have to.

I do not think a development of fricatives into affricates, outside of clusters like *ns, is well-attested at all.

[HoskhMatriarch]

about aspirated stops, if one way they can arise is through gemination, but most languages with aspirated stops and affricates don't have aspirated fricatives, what normally happens to the fricatives? Can they just turn into affricates?

They just stay geminates in most cases, I believe. /pp/ is [ppʰ] and /ss/ stays [ss], never gaining aspiration in the first place.

Another option is voicing. It's not too rare to find languages that contrast lack vs. presence of aspiration in stops, but voicing vs. voicelessness in fricatives. (Two good examples are Icelandic and Nivkh.) Often the unaspirated stops may allow voiced allophones in some positions — but they do not have to.

I do not think a development of fricatives into affricates, outside of clusters like *ns, is well-attested at all.

It's not? I thought that's the most common source of /t͡ɬ/... Then where does /t͡ɬ/ come from?

Yeah, I just came here trying to figure out how to get aspiration (I am among the few who like aspiration better than voicing), but I don't want geminate consonants, no thank you (and I don't want breathy voice either). I also don't want a [ʒ] sound, because for some odd reason I generally dislike that sound even though I like [z] and [ʃ].

[vokzhen]

I do not think a development of fricatives into affricates, outside of clusters like *ns, is well-attested at all.

It's not? I thought that's the most common source of /t͡ɬ/... Then where does /t͡ɬ/ come from?

/ll/ > [tɬ] (Icelandic, along with /nn/ > [tn̥])
*ta > /tɬa/ (Nahuatl)
/tl kl/ > [tɬ] (various)
possibly kx>kʟ̝̊>tɬ (the first stage in some African languages)
Those are the ones I'm aware of, though I'm sure there's other sources as well.

/tɬ/ doesn't generally exist without [ɬ] being present in the language, the only exception I've ever found was one dialect of Wintu. [ɬ] often arises by merger with a fricative (/hl/ in Icelandic, *sl in Welsh) or final/coda devoicing (Nahuatl, many others), but there's also geminate devoicing (some Inuit /ll/ [ɬɬ]), *r>ɬ in Forest Nenets and *s,*š>*ɬ>various outcomes in Khanty, *l>ɮ with allophone [ɬ] after voiceless consonants in some Mongolian and Northwest Caucasian, as well as I'm pretty sure quite a few other possibilities.

There does seem to be sporadic changes of fricatives into affricates under the influence of other languages, where the minority language has /ʃ/ and the majority language has /tʃ/ with no /ʃ/, and the minority has spontaneous ʃ>tʃ. Pretty sure I've heard of that happening to a few languages in Mesoamerica (under Spanish influence) and in the Sudan/Ethiopian Highlands/Uganda area, though insert disclaimer about humans and finding patterns with only two data points.

[Nortaneous]

s > θ > tθ is attested in Burmese. [ts] is a possible allophone of /s/ in Manchu.

s > ɬ happened in one Austroasiatic language, followed by c > s and the reintroduction of /c/ through loans, but *s seems to be unstable there -- Kri split final /s/ into r̥/j̥ depending on the initial, some other AA languages had s c > θ s, etc. (Was Written Burmese <c> /ts/ or /c/?) Didn't Taishanese also have s > ɬ?

[HoskhMatriarch]

I do not think a development of fricatives into affricates, outside of clusters like *ns, is well-attested at all.

It's not? I thought that's the most common source of /t͡ɬ/... Then where does /t͡ɬ/ come from?

/ll/ > [tɬ] (Icelandic, along with /nn/ > [tn̥])
*ta > /tɬa/ (Nahuatl)
/tl kl/ > [tɬ] (various)
possibly kx>kʟ̝̊>tɬ (the first stage in some African languages)
Those are the ones I'm aware of, though I'm sure there's other sources as well.

/tɬ/ doesn't generally exist without [ɬ] being present in the language, the only exception I've ever found was one dialect of Wintu. [ɬ] often arises by merger with a fricative (/hl/ in Icelandic, *sl in Welsh) or final/coda devoicing (Nahuatl, many others), but there's also geminate devoicing (some Inuit /ll/ [ɬɬ]), *r>ɬ in Forest Nenets and *s,*š>*ɬ>various outcomes in Khanty, *l>ɮ with allophone [ɬ] after voiceless consonants in some Mongolian and Northwest Caucasian, as well as I'm pretty sure quite a few other possibilities.

There does seem to be sporadic changes of fricatives into affricates under the influence of other languages, where the minority language has /ʃ/ and the majority language has /tʃ/ with no /ʃ/, and the minority has spontaneous ʃ>tʃ. Pretty sure I've heard of that happening to a few languages in Mesoamerica (under Spanish influence) and in the Sudan/Ethiopian Highlands/Uganda area, though insert disclaimer about humans and finding patterns with only two data points.

OK. I know how to get /ɬ/ fine though, that's why I was asking about the /t͡ɬ/. So if /tl/ can go to /t͡ɬ/, what leads to /t͡ɬʰ/, /ttl/?

[vokzhen]

OK. I know how to get /ɬ/ fine though, that's why I was asking about the /t͡ɬ/. So if /tl/ can go to /t͡ɬ/, what leads to /t͡ɬʰ/, /ttl/?

Not sure, but aspiration seems most likely to me. That is, a language with a /tɬ tɬʰ/ contrast probably gained aspirated after it gained /tɬ/, or at least reinforced it with a secondary source of aspiration. Or maybe shifting an entire series to lateral, like /s ts tsʰ/ > /ɬ tɬ tɬʰ/. Off the top of my head the languages I know of with /tɬ/ either don't have more than one series (at least in /tɬ/, as Nahuatl), or have multiple series as far back as we reconstruct (Na-Dene), though maybe someone else has some examples.

[Nortaneous]

tʰl

or kl kʰl > t͡ɬ t͡ɬʰ

or hr r > kʟ̥ gʟ > tɬʰ tɬ

[HoskhMatriarch]

tʰl

or kl kʰl > t͡ɬ t͡ɬʰ

or hr r > kʟ̥ gʟ > tɬʰ tɬ

Does that mean /dl/ would also have to (not possibly, but necessarily) turn into /d͡ɮ/? If so, does that mean the language would gain an /ɮ/ too?

There arent a whole lot of languages that use pharyngealized vowels, but I imagine that pharungalized vowels could turn into nasal vowels unconditionally.

I know that post was from a while back, but I really doubt that. I've never heard of that happening. IIRC, the times when languages lose pharyngealized vowels, the pharyngealized vowels turn into low vowels or back vowels, but I know they don't turn into nasalized vowels. Also, even if that could happen, it's really hard to get pharyngealized vowels compared to nasal vowels since you either have to have pharyngeal (or other radical) consonants, pharyngealized consonants, or a uvular rhotic (possibly the same sounds not patterning as a rhotic would work, but I've only heard of languages gaining pharyngealized vowels from uvular consonants that have it as a rhotic e.g. Danish, and it seems to be because of the thing where some languages have a tendency to drop rhotics in the coda and change them into other sounds), so you might as well just go for nasal vowels directly like French.

[Travis B.]

tʰl

or kl kʰl > t͡ɬ t͡ɬʰ

or hr r > kʟ̥ gʟ > tɬʰ tɬ

Does that mean /dl/ would also have to (not possibly, but necessarily) turn into /d͡ɮ/? If so, does that mean the language would gain an /ɮ/ too?

For starters, a component of affricate or a laterally-released consonant need not be an independent phoneme. So just because there may be an [dɮ] does not by any means require there to be a /ɮ/.

Secondarily, look carefully at the example. It involves devoicing of the affricates in question, probably accompanied by a generalized devoicing of plosives in general. Hence the result would [tɬ] not [dɮ]. Without devoicing, though, there would have, for /hr r/, as specified, and likely /tl dl/ or /kl gl/, been [tɬʰ dɮ], if voiceless consonants get aspirated, or [tɬ dɮ], if no aspiration takes place.

[HoskhMatriarch]

tʰl

or kl kʰl > t͡ɬ t͡ɬʰ

or hr r > kʟ̥ gʟ > tɬʰ tɬ

Does that mean /dl/ would also have to (not possibly, but necessarily) turn into /d͡ɮ/? If so, does that mean the language would gain an /ɮ/ too?

For starters, a component of affricate or a laterally-released consonant need not be an independent phoneme. So just because there may be an [dɮ] does not by any means require there to be a /ɮ/.

Secondarily, look carefully at the example. It involves devoicing of the affricates in question, probably accompanied by a generalized devoicing of plosives in general. Hence the result would [tɬ] not [dɮ]. Without devoicing, though, there would have, for /hr r/, as specified, and likely /tl dl/ or /kl gl/, been [tɬʰ dɮ], if voiceless consonants get aspirated, or [tɬ dɮ], if no aspiration takes place.

I was just inferring that since having a /t͡ɬ / seems to require an /ɬ/ that it might be the same for /d͡ɮ/ and /ɮ/. Also, besides the /hr r/ ones, there's no devoicing...

[vokzhen]

I was just inferring that since having a /t͡ɬ / seems to require an /ɬ/ that it might be the same for /d͡ɮ/ and /ɮ/. Also, besides the /hr r/ ones, there's no devoicing...

As far as I can tell /dɮ/ doesn't even exist. Most sources I've found that list /dɮ/ (actually generally they use <dl>) refer to a voiceless affricate, a handful of others have [dɮ] as an intervocally- or initially-voiced allophone of /tɬ/, and Wikipedia lists one language with it as an allophone of /ɮ/. If there's a shift of tʰl tl dl to affricates, I'd probably expect /dl/ to either assimilate [l] or [ll] rather than [dɮ]. Or maybe it would go to [dɮ] and then quickly lenite to [ɮ], following the pattern of other voiced coronals like /dz dʒ/ that deaffricate more freely than their voiceless pairs.

[HoskhMatriarch]

I was just inferring that since having a /t͡ɬ / seems to require an /ɬ/ that it might be the same for /d͡ɮ/ and /ɮ/. Also, besides the /hr r/ ones, there's no devoicing...

As far as I can tell /dɮ/ doesn't even exist. Most sources I've found that list /dɮ/ (actually generally they use <dl>) refer to a voiceless affricate, a handful of others have [dɮ] as an intervocally- or initially-voiced allophone of /tɬ/, and Wikipedia lists one language with it as an allophone of /ɮ/. If there's a shift of tʰl tl dl to affricates, I'd probably expect /dl/ to either assimilate [l] or [ll] rather than [dɮ]. Or maybe it would go to [dɮ] and then quickly lenite to [ɮ], following the pattern of other voiced coronals like /dz dʒ/ that deaffricate more freely than their voiceless pairs.

Then maybe that kind of shift to affricates doesn't even happen...

[Ketumak]

If you're wondering what the threadstarter (me) took from all this, the phonology of Modern Standard Lemohai is now published on my blog as is a note on some salient dialect features. As these posts show, I like to have a familiar core of sounds with a sprinkling of non-European-ness, rather than anything more radical.