Page 1 of 2

Arabic and Korean emphatic/tense consonants

Posted: Tue Nov 10, 2020 2:06 pm
by Kuchigakatai
From the Linguistic Miscellany Thread:
Kuchigakatai wrote: Mon Nov 09, 2020 9:06 am Arabic borrowed English "to shoot" (in the context of soccer), some time ago, as present-tense yaʃu:tˤu, past-tense ʃa:tˤa, verbal noun ʃa:tˤ, which hilariously matches both the English inflection and related noun (shoot, he shot, a shot) all while using a perfectly native inflectional pattern found in triconsonantal roots with /w/ as the second consonant (e.g. yaqu:lu 'he speaks', qa:la 'he spoke', which has the root q-w-l).
Zju wrote: Mon Nov 09, 2020 4:00 pmShot is not the first thing I thought of when I read past-tense ʃa:tˤa, verbal noun ʃa:tˤ...
Kuchigakatai wrote: Mon Nov 09, 2020 5:49 pmI should've mentioned those are Standard Arabic phonemes. In phonetic reality they're more like [ˈʃɒːtˠˤɒ] and [ʃɒːtˠˤ].
Zju wrote: Tue Nov 10, 2020 12:34 pmNow I'm intrigued - is /tˤ/ really [tˠˤ] as opposed to [tˠ] or [tˤ]? How common is that intralanguagewise, and does it occur in other languages?
I think this quote from Mark Cowell's A Reference Grammar of Syrian Arabic (1964: 6-7), which is one of the best presentations of the "emphatic" consonant contrast in Arabic I've seen, might help:
VELARIZATION1: ṭ, ḍ, ṣ, ẓ, ḍ, ḅ, ṃ, ṇ, ḷ, ṛ, (?).

The dot under these letters represents a "heavy" resonance which is the effect of relatively low-pitched concentrations of acoustic energy — in contrast to the "thin" or "light" quality of the sounds transcribed without the dot. (Note that ḥ [p. 4] is not one of the velarized sounds; its dot is merely to distinguish it from h.)

In producing the plain sounds (i.e. those transcribed without the dot), the tongue is usually arched upward and forward into a single hump (in profile), leaving the pharyngeal and velar passages relatively open. For the velarized sounds, on the other hand, the profile of the tongue usually tends to be two-humped and low in the middle; the back hump narrows the velar and pharyngeal passages.

The lips may also play a part in the produce the heavy resonance; velarization is sometimes accompanied by protrusion and pursing of the lips, while retraction and spreading of the lips help make the lighter, thinner resonance.

Examples of the contrast between plain and velarized sounds:
Plain .......... Velarized
tīn ‘figs’ ..... ṭīn ‘mud’2
[9 more examples omitted]

Speakers of English and many other languages are apt to be more sensitive to the effects of velarization on contiguous vowels than to the differences between plain and velarized consonants themselves. Compare dall ‘to indicate’ with ḍaḷl ‘to remain’, sədd ‘close, block’ with ṣəḍḍ ‘repulse, refuse’. [pp. 10, 11]

Velarization is usually not limited to a single sound in a word, but commonly affects whole syllables and often hwole words: ḍaḷḷ, ṃaḅṣū´ṭ, ẓā´ḅeṭ.

The dental obstruents t/ṭ, d/ḍ, s/ṣ, and z/ẓ are the only ones of these pairs that differentiate many words independently as illustrated above. With the others, the distinction between plain and velarized is usually a variation conditioned by the neighboring sounds, and is potentially significant only next to the vowel a and in the absent of dental obstruents.

Since velarization mainly affects sound sequences that involve dental obstruents, these obstruents are taken as the focal points of velarization wherever possible. Our transcription regularly shows velarization for these sounds, but not for the other kinds of sounds affected in their neighborhood. Thus in the word bə´ṭlaع, for example, the dot under the ‘t’ implies that the b, the ə, and the l are normally also velarized.

This economical use of subscript dots is not unambiguous, since the scope of velarization — the "neighborhood" of a dotted letter — has not been defined, nor is there, apparently, any simple way to define it. In fact the scope of velarization varies considerably from word to word, speaker to speaker, and region to region. Furthermore, the velarization may vary in intensity; some parts of a word may be strongly velarized, other parts weakly.

———

1 The term ‘velarization’ is not altogether satisfactory as a name for this phonological component. Note that the post-velar sounds x, ġ, and q are not inherently "velarized"; they may be either "plain" or "velarized", depending on the neighboring sounds. The term ‘pharyngealization’, which has sometimes been used instead of ‘velarization’, is even more misleading, since the pharyngeal spirants ḥ and ع have still less in common with the velarized sounds than the post-velars have.

Evidently the air-stream turbulence produced by primary velar or pharyngeal stricture has sound effects quite unrelated — in Arabic, at lesat — to the efect of so-called secondary stricture in these passages. The secondary stricture does not produce audible turbulence, but serves to modify the resonating chamber.

The traditional term ‘emphatic’ is also a bad name for the velarized sounds, since it suggests (erroneously, it would seem) that these sounds are more forcefully or tensely articulated than the plain sounds.

2 Velarized ṭ is usually unaspirated while plain t is somewhat aspirated.
Basically, secondary articulation is involved, and the whole of the back of the tongue is raised backwards to add constriction and so modulate the sound of the consonant.

As the text says, there's a small element of protruded rounding too, which I tried to render above with the rounded back vowel [ɒ], and which in particularly narrow transcriptions I have on occasion seen rendered as e.g. [tˠˤʷ]. I don't personally perceive this roundedness much, but apparently Swahili speakers do, and you can see this in the way they borrow Arabic words with emphatic consonants using their native /w/ sometimes: swala 'prayer' < Arabic صلاة‎ sˤala (also pronounced sˤalaːt, stressed on the second syllable) 'prayer, Salah/Salat'. Not that Swahili speakers are consistent about it: sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa, safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.

I have some things to say about the similarly interesting Korean tense consonants, with links to sound samples, but that'll happen in another later post.

Re: Arabic and Korean emphatic/tense consonants

Posted: Thu Nov 12, 2020 10:35 pm
by Nortaneous
My impression is that "velarization" and "pharyngealization", while possibly phonetically real, aren't contrastively operative - +ARGH vowels in Sino-Tibetan are described alternately as "velarized", "uvularized", "pharyngealized", or "tense", but it's not clear that they aren't in some sense the same thing. They can still be phonologically operative, though - IIRC Xun Gong says somewhere that uvularized vowels in ST agree in uvularization with initials, but velarized vowels don't - so /kaˠ qaˠ/, but /*kaʶ qaʶ/.

Then again, the only (AFAIK) attestation of a contrast between velarization and pharyngealization is in Zbu Rgyalrong: /aˠ/ [ɑ] vs. /oʁ/ [ɑˤ].

Re: Arabic and Korean emphatic/tense consonants

Posted: Thu Nov 12, 2020 11:27 pm
by bradrn
Kuchigakatai wrote: Tue Nov 10, 2020 2:06 pm Not that Swahili speakers are consistent about it: sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa, safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.
This is interesting to me, since what I presume to be the Hebrew cognate — צדקה /tsədaka/ — shares its meaning with the Swahili word rather than with Arabic. Is this a case where a word has undergone exactly the same historical development in two widely separated languages (which seems unlikely to me), or is there something else happening here?
I have some things to say about the similarly interesting Korean tense consonants, with links to sound samples, but that'll happen in another later post.
Yes please! I’ve been wondering about those for quite a while.
Nortaneous wrote: Thu Nov 12, 2020 10:35 pm +ARGH vowels in Sino-Tibetan
???
Then again, the only (AFAIK) attestation of a contrast between velarization and pharyngealization is in Zbu Rgyalrong: /aˠ/ [ɑ] vs. /oʁ/ [(w)ɑˤ].
That paper you linked earlier about Showu/Zbu only mentioned velarisation — where did you read about pharyngealisation?

Re: Arabic and Korean emphatic/tense consonants

Posted: Thu Nov 12, 2020 11:51 pm
by Kuchigakatai
bradrn wrote: Thu Nov 12, 2020 11:27 pm
Kuchigakatai wrote: Tue Nov 10, 2020 2:06 pm Not that Swahili speakers are consistent about it: sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa, safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.
This is interesting to me, since what I presume to be the Hebrew cognate — צדקה /tsədaka/ — shares its meaning with the Swahili word rather than with Arabic. Is this a case where a word has undergone exactly the same historical development in two widely separated languages (which seems unlikely to me), or is there something else happening here?
I think you misread what I wrote, due to my own poor writing. I should've used a list instead, and maybe repeat the gloss of sˤadaqa, like this:

Not that Swahili speakers are consistent about it:
- sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa 'alms, charity'
- safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.
Nortaneous wrote: Thu Nov 12, 2020 10:35 pm +ARGH vowels in Sino-Tibetan
???
This is a long-running joke of Nortaneous, using "[±argh]" as a phonological feature, as you would use [+strident] or [+voiced]. It's supposed to mean some sort of variation of more standard vowels, involving something vague around lax/-ATR, creaky voice, velaro-pharyngealization... While funny, it is ironically useful because works on phonology and phonetics may often disagree a lot about the features involved in two vowel series in a given specific language, while agreeing there are two vowel series involved (say, /i e a o/ vs. /iʶ eʶ aʶ oʶ/ as phonemes, or whatever).

Re: Arabic and Korean emphatic/tense consonants

Posted: Fri Nov 13, 2020 12:55 am
by bradrn
Kuchigakatai wrote: Thu Nov 12, 2020 11:51 pm
bradrn wrote: Thu Nov 12, 2020 11:27 pm
Kuchigakatai wrote: Tue Nov 10, 2020 2:06 pm Not that Swahili speakers are consistent about it: sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa, safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.
This is interesting to me, since what I presume to be the Hebrew cognate — צדקה /tsədaka/ — shares its meaning with the Swahili word rather than with Arabic. Is this a case where a word has undergone exactly the same historical development in two widely separated languages (which seems unlikely to me), or is there something else happening here?
I think you misread what I wrote, due to my own poor writing. I should've used a list instead, and maybe repeat the gloss of sˤadaqa, like this:

Not that Swahili speakers are consistent about it:
- sadaka 'alms, charity' < Arabic صدقة‎ sˤadaqa 'alms, charity'
- safi 'clean, not dirty' < Arabic صافٍ sˤaːfin (colloquial [ˈsˤɒːfi]) 'pure, clear, bright'.
Oops, I did indeed misread that as “sˤadaqa 'clean, not dirty'”. Sorry!

Re: Arabic and Korean emphatic/tense consonants

Posted: Fri Nov 13, 2020 8:55 am
by Nortaneous
bradrn wrote: Thu Nov 12, 2020 11:27 pm
Nortaneous wrote: Thu Nov 12, 2020 10:35 pm +ARGH vowels in Sino-Tibetan
???
most notably Nuosu yr ur
That paper you linked earlier about Showu/Zbu only mentioned velarisation — where did you read about pharyngealisation?
Xun Gong's grammar:
La coda -ʁ a la distibution la plus restreinte parmi toutes les codas : elle ne se trouve qu’après les voyelles ɐ et o. Comme il est typique pour les uvulaires dans cette langue, elle est une uvulaire fortement pharyngalisée, avec l’épiglotte à position très basse. On constate plus souvent une réalisation réduite, comme pharyngale -ʕ ou une nuance de pharyngalisation ˤ.

Ce caractère pharyngal pousse la qualité vocalique fortement vers la direction de æ. Pour les locuteurs âgés, -ɐʁ se prononce avec une pharyngalisation vocalique d’ampleur uniforme [æˤ], ou avec un accroissement progressif de la pharyngalisation : [æʕ] ou [æ͡æˤ]. Pour -oʁ, la pharyngalisation a poussé o vers ɒ ∼ ɑ], donc variablement [ɑˤ], [ɒ͡ʕ] ou [ɒ͡ɑˤ].

L’analyse en tant que -oʁ est adoptée dans cette étude est partiellement basée sur la comparaison dans d’autres dialectes du zbu et d’autres langues rgyalrongs. On note néanmoins le même effet de pharyngalisation qui relie ɐ à [æʕ] et o à [ɒʕ].

Entre les générations présentes, un changement phonétique a eu lieu, de sorte que -ɐʁ et -oʁ se sont confondues. Chez les locuteurs du type innovateur, donc, la poussée vers æ a été menée à ton terme : toutes les voyelles devant -ʁ sont devenues æ. Celles qui sont transcrites -ɐʁ et -oʁ dans cette thèse sont toutes les deux prononcées comme -ɐʁ dans le type conservateur, surtout dans la réalisation monophtongue [æˤ].
Pharyngealization has also been described for Hongyan Qiang.

Re: Arabic and Korean emphatic/tense consonants

Posted: Fri Nov 13, 2020 11:44 am
by Hallow XIII
That last paragraph is fairly interesting:
Xun Gong wrote:Entre les générations présentes, un changement phonétique a eu lieu, de sorte que -ɐʁ et -oʁ se sont confondues. Chez les locuteurs du type innovateur, donc, la poussée vers æ a été menée à ton terme : toutes les voyelles devant -ʁ sont devenues æ. Celles qui sont transcrites -ɐʁ et -oʁ dans cette thèse sont toutes les deux prononcées comme -ɐʁ dans le type conservateur, surtout dans la réalisation monophtongue [æˤ].
Isn't this essentially what (is theorized to have) happened in Akkadian?

Re: Arabic and Korean emphatic/tense consonants

Posted: Fri Nov 13, 2020 8:05 pm
by Nortaneous
More detail on +ARGH vowels:

Nuosu has a vowel system of /a̠ ɛ̠ ɔ̠ ɿ̠ v̠̩ i ɯ o ɿ v̩/, which could alternately be analyzed as /i ɯ u ɿ v̩/ + [±ARGH]. /ɿ̠ v̠̩/ can only occur in the mid tone (and the rising tone, which is a recent development from the mid tone), and are the only vowels to which this applies.

Bai has a lot of dialects which are described in a lot of different ways, but the register system involves "tense voice" (transcribed with an underline), and in some dialects also "harsh voice". I'm not sure where Wikipedia's claim that Bai also has strident vowels comes from.

Hongyan Qiang has /a ə i u/ plus length, rhotacization, and pharyngealization. Other Qiang dialects probably have ±ARGH contrasts that just haven't been noticed yet, and were instead described as vowel place contrasts.

Zbu Rgyalrong has contrastive vowel velarization, in addition to two pharyngealized vowels. This is almost certainly a conservative feature. Japhug had *Vˠ > ɣV (or ŋV / N_) IIRC, and I don't know what happened in Situ (lost without compensation??) or Tshobdun.

Khroskyabs retains velarized vowels in the Phosul (= Puxi?) dialect; in the other dialects (which form a branch), it's lost, but *o *oˠ > e o. I don't know what 'Jorogs does. Stau-Dgebshes doesn't have velarized vowels AFAIK.

Lhagang Choyu has velarized vowels.

Nyagrong Minyag has five uvularized vowels, /iʶ yʶ uʶ əʶ aʶ/, and seven plain vowels, /i y u ɛ ə ɔ a/.

Chukotkan (probably Chukotko-Kamchatkan in general but Itelmen is hard) can be reconstructed with an [±ARGH] contrast on vowels which developed into the dominant/recessive vowel harmony. This is maybe a coincidence?

Sahaptian had dominant-recessive harmony with strong /ɑ o i ɨ/ and weak /æ u i ɨ/. This can obviously be connected to Chukotkan - maybe also Salishan, which I know nothing about.

Middle Korean had height harmony between /a ʌ o/ and /e ɨ u/, but AFAIK this wasn't dominant-recessive and was strictly determined by the initial syllable, as in Finnish and Turkish. I could be wrong about this tho. If we accept tongue root or height harmony as adjacent to vocalic [±ARGH], this also adds Mongolic and, although I'd have to check to be sure, maybe Kusunda.

Re: Arabic and Korean emphatic/tense consonants

Posted: Sat Nov 14, 2020 2:27 am
by Creyeditor
Nortaneous wrote: Fri Nov 13, 2020 8:05 pm Bai has a lot of dialects which are described in a lot of different ways, but the register system involves "tense voice" (transcribed with an underline), and in some dialects also "harsh voice". I'm not sure where Wikipedia's claim that Bai also has strident vowels comes from.
I think Wikipedia uses harsh voice and strident vowels as quasi-synonyms. At least in the article on strident vowels.

Re: Arabic and Korean emphatic/tense consonants

Posted: Sat Nov 14, 2020 4:39 am
by Moose-tache
I clicked on this thread because it said Korean, and now I feel like the guy who showed up to an Amway meeting because he thought there would be pie.

Re: Arabic and Korean emphatic/tense consonants

Posted: Sat Nov 14, 2020 4:55 am
by Raphael
Moose-tache wrote: Sat Nov 14, 2020 4:39 am I clicked on this thread because it said Korean, and now I feel like the guy who showed up to an Amway meeting because he thought there would be pie.
When I first saw the title of this thread, I thought it was about some ridiculous theory proposing a relationship between Arabic and Korean, and I thought, "Huh? Kuchigakatai doesn't strike me as the kind of person who'd fall for that kind of thing..."

Re: Arabic and Korean emphatic/tense consonants

Posted: Sat Nov 14, 2020 3:32 pm
by dhok
Old Chinese had an inventory of /a e o i ɨ u/ with some sort of distinction that may have been pharyngealization. Baxter and Sagart reconstruct it for the consonants, except that then you get e.g. a distinction between *q and *qˤ, which seems unlikely. Based on areal features it seems at least plausible that it actually belonged on the vowel.

Re: Arabic and Korean emphatic/tense consonants

Posted: Sat Nov 14, 2020 5:59 pm
by Nortaneous
dhok wrote: Sat Nov 14, 2020 3:32 pm Old Chinese had an inventory of /a e o i ɨ u/ with some sort of distinction that may have been pharyngealization. Baxter and Sagart reconstruct it for the consonants, except that then you get e.g. a distinction between *q and *qˤ, which seems unlikely. Based on areal features it seems at least plausible that it actually belonged on the vowel.
IIRC the main argument that it wasn't a vowel feature was that it was disregarded in rhymes. Contrastive uvularization on pharyngeals is attested, but the only attestation I know of outside NWC and NEC is the marginal and unsystematic ʁˤ in Zwara Berber, and I don't know of any cases of contrastive consonant pharyngealization in Sino-Tibetan at all.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 12:48 am
by Moose-tache
IIUC, the earliest rhyme tables in Chinese were written centuries after the Old Chinese period. And even conscientious writers ignore a lot of detail when describing their own language (Sequoyah seemed pretty cavalier about Cherokee consonant clusters). Is there any strong evidence that phonation couldn't be ignored in rhyme tables?

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 3:36 am
by dhok
Rhyme tables aren't the same as rhyme evidence. The earliest Chinese poetry, e.g. in the Shijing (earliest poems from about 1000 BC), had end-rhyme like poetry in modern European languages. The first rhyme dictionary was compiled in the Han period, and survived until at least the mid-Song, but is now lost except for fragments. (Maybe a copy will turn up someday--China is a big place.)

You're correct that the Qièyùn dates from 601 AD, and is the earliest complete rhyme dictionary we have directly, but is for Middle Chinese.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 6:55 am
by zompist
dhok wrote: Mon Nov 16, 2020 3:36 am Rhyme tables aren't the same as rhyme evidence. The earliest Chinese poetry, e.g. in the Shijing (earliest poems from about 1000 BC), had end-rhyme like poetry in modern European languages. The first rhyme dictionary was compiled in the Han period, and survived until at least the mid-Song, but is now lost except for fragments. (Maybe a copy will turn up someday--China is a big place.)
There's one more bit of rhyme evidence: the writing system itself. It was noticed (by Duàn Yùcái, d. 1815) that characters which shared a phonetic component could rhyme in the Shījīng system. This provided a greatly expanded dataset, as scholars were no longer restricted to words that happened to be used in the Shījīng.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 9:19 am
by Moose-tache
I get all that. What I meant was, and I recognize that I don't know much about historical Chinese phonology so bear with me, is it possible for phonemic features to exist that simply are not relevant to rhymes?

For example, in modern pop music assonance with a final nasal is almost always treated like a full rhyme. Someone like Katie Perry or Drake will treat "time" and "fine" as no different than "time" and "chime." If all we had to go by was early 21st century American pop music, we could easily conclude that English had only one coda nasal, like Japanese. So how do we know that Old Chinese speakers didn't make phonemic distinctions that were simply ignored for the purposes of rhyme?

As for the phonetic component of characters, we already know that this system cuts a few corners, right? Not every phonemic distinction in the language is clearly laid out through phonemic components, or so I understand.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 12:44 pm
by Zju
What was the topic again? I'd be grateful if someone knowledgable weighed in on the topic of Korean tense consonants. The most I've seen is 'Korean tense consonants are sometimes said to be this' and 'Korean tense consonants are described by some as that'.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 2:40 pm
by Creyeditor
I think different Korean sociolects and dialects differ. I heard phonologists claiming that in older speakers the following vowel is glottalized and in younger speakers it has evolved into a pitch difference.

Re: Arabic and Korean emphatic/tense consonants

Posted: Mon Nov 16, 2020 3:01 pm
by dɮ the phoneme
Creyeditor wrote: Mon Nov 16, 2020 2:40 pm I think different Korean sociolects and dialects differ. I heard phonologists claiming that in older speakers the following vowel is glottalized and in younger speakers it has evolved into a pitch difference.
If you're referring to the tonogenesis currently taking place in Seoul Korean, the distinction that's getting transphonologized as tone is the aspirated vs. plain stop contrast, not the fortis vs. lenis one.

According to this paper:
[A]spirated stops have the longest VOT values, fortis stops have the shortest values, and lenis stops have intermediate values; F0 on the following vowel is higher for aspirated and fortis stops than for lenis stops; vowels following aspirated and lenis stops have breathier voice quality than vowels following fortis stops as indicated by higher H1-H2 values

[...]

These studies show that Seoul Korean is in fact in the process of losing the VOT (Voice Onset Time) contrast between aspirated and lenis stops (/ph th kh/ vs. /p t k/) in phrase-initial position and the formerly redundant pitch difference—a high pitch on vowels following aspirated stops and a low
over time—is replacing the VOT difference as the primary cue of the contrast. A similar change is reported for a dialect of Chinese Korean