Palatalization of ejectives

[The Hanged Man]

Is palatalization of ejectives possible? If so, what outcome would it have?

[Morrígan]

Well, you would end up with palatalized ejectives.

Remember, palatalization is a secondary articulation; ejection (?), like aspiration, relates the the state of the glottis, and is an phonation type.

[Aurora Rossa]

I would imagine so. I can't think of any reason they wouldn't palatalize under the same conditions as regular stops. I don't think they would differ much in outcome, though maybe some acoustic factors could come into play.

[chris_notts]

A lot of Salishan languages have ejective prevelars, which are pretty close to k_j_>. You could also look at the languages of the Caucasus, which tend to be a hotbed for both secondary articulations and ejective consonants.

[sangi39]

A lot of Salishan languages have ejective prevelars, which are pretty close to k_j_>. You could also look at the languages of the Caucasus, which tend to be a hotbed for both secondary articulations and ejective consonants.

Yep, Ubykh had them at velar and uvular POAs, as does Abkhaz.

[roninbodhisattva]

Yes. It is.

[nebula wind phone]

Mam has /kʲʼ/. There's another natlang example.

Or do you mean "Is it possible to have a sound change in which every ejective in the language becomes palatalized?" Because that would be odd. Don't know of any languages where it's happened, and I wouldn't expect to find any.

[Pabappa]

I know this is a bit of astrochironecromancy, but I just came across this as it was literally the only googlt hit for "palatalization of ejectives". I think that /k_>/ is a bit more likely to be further back than other velars although Im not sure why. This could prevent it from being palatalized since /q/ cant be palatalized.* This tendency could then carry over to any other ejectives in the language. Also, I think ejectives can change into pharyngealized consonants, which in turn I think have a tendency to be paired with back vowels, meaning that it would kill palatalization. However, if you use that route, you wouldnt have any ejectives anymore unless you somehow switched them back.

I dont know how you;d poronounce /q_j/, because supposedly it is impossible, but a few languages seem to have it.

[vokzhen]

I think that /k_>/ is a bit more likely to be further back than other velars although Im not sure why.

Do you have some examples? I know Chechen and Ingush have a postvelar /k'/ and a backer-than-normal-uvular /q'/ but I just assumed that was a quirk of the language.

Also, I think ejectives can change into pharyngealized consonants, which in turn I think have a tendency to be paired with back vowels, meaning that it would kill palatalization. However, if you use that route, you wouldnt have any ejectives anymore unless you somehow switched them back.

Pretty sure that while they can change into pharyngealized consonants, it's not really anything to do with an inherent back-ness of ejectives. Rather, the sound /t'/ [tˀ] is just a slightly shift of POA away from [tˁ], which happened to happen in Semitic and Circassian. I don't know what happens in Circassian to palatal(ized) ejectives, Wikipedia implies that only some ejectives have pharyngealized realizations and doesn't include the palatalized ones in the examples so it may be they resist it, but the grammars I have don't go into enough detail to confirm or contradict that.

This could prevent it from being palatalized since /q/ cant be palatalized.*
I dont know how you;d poronounce /q_j/, because supposedly it is impossible, but a few languages seem to have it.

I don't think it's impossible - at least I think I can pronounce [qʲ]. And as you said, it's at least present in some Northwest Caucasian languages, and interestingly includes Abaza which, according to Wikipedia, has a full vowel inventory /i e u o a ə/ rather than the vertical system of the other NWC languages.

[Zaarin]

I think that /k_>/ is a bit more likely to be further back than other velars although Im not sure why.

Do you have some examples? I know Chechen and Ingush have a postvelar /k'/ and a backer-than-normal-uvular /q'/ but I just assumed that was a quirk of the language.

I suppose there's also the Semitic/Egyptian (and Berber?) shift of /k'/ > /q/, which I think happened in a few other languages. But I don't think it's any kind of universal that ejective k is further back than non-ejective k.

[Αυτοβοτα]

I think that /k_>/ is a bit more likely to be further back than other velars although Im not sure why.

Do you have some examples? I know Chechen and Ingush have a postvelar /k'/ and a backer-than-normal-uvular /q'/ but I just assumed that was a quirk of the language.

I suppose there's also the Semitic/Egyptian (and Berber?) shift of /k'/ > /q/, which I think happened in a few other languages. But I don't think it's any kind of universal that ejective k is further back than non-ejective k.

That's with a concurrent shift of all ejectives/glottalics being realized with pharyngeal coarticulation, though. /q/ makes more sense as /kˤ/ instead of /kʼ/, especially along with /tˤ (t)sˤ (t)ɬˤ/ etc.

[Zaarin]

I think that /k_>/ is a bit more likely to be further back than other velars although Im not sure why.

Do you have some examples? I know Chechen and Ingush have a postvelar /k'/ and a backer-than-normal-uvular /q'/ but I just assumed that was a quirk of the language.

I suppose there's also the Semitic/Egyptian (and Berber?) shift of /k'/ > /q/, which I think happened in a few other languages. But I don't think it's any kind of universal that ejective k is further back than non-ejective k.

That's with a concurrent shift of all ejectives/glottalics being realized with pharyngeal coarticulation, though. /q/ makes more sense as /kˤ/ instead of /kʼ/, especially along with /tˤ (t)sˤ (t)ɬˤ/ etc.

Fair, but the shift also occurred in the Semitic languages that retained ejectives, like Akkadian and Biblical Hebrew, as well as Egyptian which is generally agreed to have had ejectives.

[vokzhen]

Fair, but the shift also occurred in the Semitic languages that retained ejectives, like Akkadian and Biblical Hebrew, as well as Egyptian which is generally agreed to have had ejectives.

Sources? Everything I've seen either has emphatics as all pharyngealized (mostly older, less rigorous sources) or all ejectives, except that <q> is usually used to represent /k'/.

[Xephyr]

Fair, but the shift also occurred in the Semitic languages that retained ejectives, like Akkadian and Biblical Hebrew, as well as Egyptian which is generally agreed to have had ejectives.

Sources? Everything I've seen either has emphatics as all pharyngealized (mostly older, less rigorous sources) or all ejectives, except that <q> is usually used to represent /k'/.

If the Proto-Semitic forms were ejectives, though, then at some point they must have become pharyngealized in Arabic (and iirc Aramaic), even if the attested periods of the language only have either all-ejectives or all-pharyngealizeds.

[vokzhen]

Fair, but the shift also occurred in the Semitic languages that retained ejectives, like Akkadian and Biblical Hebrew, as well as Egyptian which is generally agreed to have had ejectives.

Sources? Everything I've seen either has emphatics as all pharyngealized (mostly older, less rigorous sources) or all ejectives, except that <q> is usually used to represent /k'/.

If the Proto-Semitic forms were ejectives, though, then at some point they must have become pharyngealized in Arabic (and iirc Aramaic), even if the attested periods of the language only have either all-ejectives or all-pharyngealizeds.

My question was more specifically about having an inventory of /t' s' q/ (or whatever), at least for longer than a short period of transition in the ejective>pharyngealized shift (and evidence from Modern South Arabian seems to point at a simultaneous pharyngealization across the the inventory, rather than step-by-step).

[Zaarin]

Fair, but the shift also occurred in the Semitic languages that retained ejectives, like Akkadian and Biblical Hebrew, as well as Egyptian which is generally agreed to have had ejectives.

Sources? Everything I've seen either has emphatics as all pharyngealized (mostly older, less rigorous sources) or all ejectives, except that <q> is usually used to represent /k'/.

If the Proto-Semitic forms were ejectives, though, then at some point they must have become pharyngealized in Arabic (and iirc Aramaic), even if the attested periods of the language only have either all-ejectives or all-pharyngealizeds.

My question was more specifically about having an inventory of /t' s' q/ (or whatever), at least for longer than a short period of transition in the ejective>pharyngealized shift (and evidence from Modern South Arabian seems to point at a simultaneous pharyngealization across the the inventory, rather than step-by-step).

You're right, I forget that <q> is often favored over <ḳ> in Semiticist transliterations for the dorsal emphatic, even when the underlying value may not be [q].

[Nortaneous]

there are some languages where k_> is backer than k. there are also some languages where k is backer than g. some papuan languages are analyzed as having a stop system of /p t q b d g/. also how did those cushitic languages get retroflexes, didn't they go t_> -> d_< -> d` or something

anyway of course ejectives can palatalize

[Tropylium]

This could prevent it from being palatalized since /q/ cant be palatalized.*
I dont know how you;d poronounce /q_j/, because supposedly it is impossible, but a few languages seem to have it.

I don't think it's impossible - at least I think I can pronounce [qʲ]. And as you said, it's at least present in some Northwest Caucasian languages, and interestingly includes Abaza which, according to Wikipedia, has a full vowel inventory /i e u o a ə/ rather than the vertical system of the other NWC languages.

Bert Vaux in A Note on Pharyngeal Features sort of proposes that "uvulars" are in fact several different types of consonants:

1. [-ATR] [-high] [+back] uvulars proper
2. [-ATR] [+high] [-back] "palatalized uvulars" (≈ palatalized pharyngealized velars)
3. [+back], non-ATR-specified "uvulars" that can even be [-high] (≈ postvelars)

He also posits that [-ATR] is not the same thing as [+RTR] though, which is what regular pharyngealized consonants would be.

(He additionally claims that voiced stops are [+ATR], and therefore voiced uvular stops of the first two types are impossible; all attested cases of /ɢ/ should be of the 3rd type.)

[Nortaneous]

Ahtena (UPSID) Ahtena Kari, James and Buck, Mildred 1975; Kari, James 1979
Awngi (SPA) Awngi Hetzron, Robert 1969
Awngi (UPSID) Awngi Hetzron, Robert 1969
Eyak (UPSID) Eyak Krauss, Michael E. 1965
Gilyak (SPA) Gilyak Panfilov, V. Z. 1968; Panfilov, V. Z. 1962
Ingush (PH) Ingush Nichols, Johanna 1996
Kadiweu (SAPHON) Kadiweu Sandalo, Filomena 1997
Kirghiz (UPSID) Kirghiz Junusaliev, B. M. 1966; Herbert, Raymond J. and Poppe, Nicholas 1963
Klamath-Modoc (UPSID) Klamath-Modoc Barker, M. A. R. 1964
Kunimaipa (SPA) Kunimaipa Pence, Alan 1966
Kunimaipa (UPSID) Kunimaipa Pence, Alan 1966
Kusunda (PH) Kusunda Watters, David E. 2006
Kwakiutl (SPA) Kwakiutl Newman, Stanley 1950; Boas, Franz 1947
Kwakiutl (UPSID) Kwakiutl Newman, Stanley 1950; Boas, Franz 1911; Grubb, D. M. 1977; Boas, Franz 1947
Lak (SPA) Lak Murkelinskij, G. B. 1967; Khajdakov, S. M. 1966; Zhirkov, Lev I. 1955
Lak (UPSID) Lak Khajdakov, S. M. 1966; Murkelinskij, G. B. 1967; Zhirkov, Lev I. 1955
Nisga'a (PH) Nisga'a Tarpent, Marie-Lucie 1987
Nuclear Tsimshian (UPSID) Nuclear Tsimshian Dunn, John Asher 1978; Dunn, John Asher 1979; Hoard, J. E. 1978; Mulder, Jean Gail 1988
Qimant (PH) Qimant Hetzron, Robert 1969
Rutul (UPSID) Rutul Dzhejranishvili, E. F. 1967; Ibragimov, Garun K. 1978
Santiago del Estero Quichua (SAPHON) Santiago del Estero Quichua Alderetes, Jorge 2001
Sauria Paharia (RA) Sauria Paharia Das, Sisirkumar 1973
Somali (SPA) Somali Armstrong, Lilias E. 1964; Andrzejewsky, B. W. 1956; Andrzejewsky, B. W. 1955
Somali (UPSID) Somali Andrzejewsky, B. W. 1956; Armstrong, Lilias E. 1964; Cardona, G. R. 1981; Farnetani, E. 1981; Andrzejewsky, B. W. 1955
Tehuelche (SAPHON) Tehuelche Adelaar, Willem and Pieter Muysken 2004
Tlingit (UPSID) Tlingit Swanton, John R. 1911; Swanton, John R. 1909; Story, Gillian L. and Naish, Constance M. 1973
Tu (UPSID) Tu Todaeva, B. X. 1973
Vilela (SAPHON) Vilela Viegas Barros, J. Pedro 2001
Western Farsi (SPA) Western Farsi Obolensky, Serge and Panah, Kambiz Yazdan and Nouri, Fereidoun Khaje 1963
Western Farsi (UPSID) Western Farsi Obolensky, Serge and Panah, Kambiz Yazdan and Nouri, Fereidoun Khaje 1963
!Xoo (GM) !Xoo Traill, Anthony 2009

Are these all postvelar? (Do they even all exist? I don't trust UPSID.)

[Zaarin]

The only two of those I know well enough to answer are Tlingit and Kwak'wala (aka "Kwakiutl"), but I've never heard either language's uvular consonants described as post-velar. Though I don't know it as well, my instinct is to say the same of Nisga'a (though I'm more familiar with Coast Tsimshian [which is probably what they mean by "Nuclear Tsimshian"?], which I also understand to be uvular).

[Nortaneous]

1. [-ATR] [-high] [+back] uvulars proper
2. [-ATR] [+high] [-back] "palatalized uvulars" (≈ palatalized pharyngealized velars)
3. [+back], non-ATR-specified "uvulars" that can even be [-high] (≈ postvelars)

He also posits that [-ATR] is not the same thing as [+RTR] though, which is what regular pharyngealized consonants would be.

Ubykh: /kʲ k kʷ qʲ q qʷ qˁ qˁʷ/. How would this deal with that?

[Sumelic]

1. [-ATR] [-high] [+back] uvulars proper
2. [-ATR] [+high] [-back] "palatalized uvulars" (≈ palatalized pharyngealized velars)
3. [+back], non-ATR-specified "uvulars" that can even be [-high] (≈ postvelars)

He also posits that [-ATR] is not the same thing as [+RTR] though, which is what regular pharyngealized consonants would be.

Ubykh: /kʲ k kʷ qʲ q qʷ qˁ qˁʷ/. How would this deal with that?

It seems the first set are velars (with possible palatalized ([+high]) and labialized secondary articulations), the second set are [-back][-atr][-rtr] uvulars (with possible palatalized ([+high]) and labialized secondary articulations), and the third set are [-atr][+rtr] uvulars. (I didn't see any discussion of labialization in the linked paper.)

[Tropylium]

1. [-ATR] [-high] [+back] uvulars proper
2. [-ATR] [+high] [-back] "palatalized uvulars" (≈ palatalized pharyngealized velars)
3. [+back], non-ATR-specified "uvulars" that can even be [-high] (≈ postvelars)

He also posits that [-ATR] is not the same thing as [+RTR] though, which is what regular pharyngealized consonants would be.

Ubykh: /kʲ k kʷ qʲ q qʷ qˁ qˁʷ/. How would this deal with that?

It seems the first set are velars (with possible palatalized ([+high]) and labialized secondary articulations), the second set are [-back][-atr][-rtr] uvulars (with possible palatalized ([+high]) and labialized secondary articulations), and the third set are [-atr][+rtr] uvulars. (I didn't see any discussion of labialization in the linked paper.)

We can largely ignore labialization yes (which would be its own feature entirely), but Vaux' approach does not accept any [+velar] feature, so we cannot discard the first two from analysis. He does not comment on plain velar/palatalized velar contrasts, but apparently the analysis would be:

Code: Select all

      kʲ k  qʲ q  qˤ
high  +  +  +  -  -
back  -  0  -  +  +
ATR   0  0  -  -  -
RTR   -? -? -  -  +

i.e. it it is necessary to treat features as ternary: any feature can be either present, absent, or unspecified.