Turtle this
« previous post | next post »
You may or may not have heard of Kucha. For those who are interested in Tocharian or Uyghur, you almost certainly would be well aware of this oasis city on the northern rim of the Taklamakan Desert in the Tarim Basin of Eastern Central Asia.
Coordinates: 41°42′56″N 82°55′56″E
Kucha is the historical seat of so-called Tocharian B, i.e., Kuśiññe Kantwo, the home of the renowned Buddhist translator, Kumārajīva (344-413), and an important center of Uyghur history and culture from the 7th to 13th centuries.
The oldest Chinese name I know of for Kucha is Qiūcí 龜茲. I recall when I first encountered it how puzzled I was. Nearly everyone without specialized knowledge who has occasion to read this name in Modern Standard Mandarin (MSM) is apt to pronounce it guīzī, but they will be wrong on both syllables.
龜 is a fun character to behold, but it is also one of the hardest characters to write. It only has 17 strokes, but they are difficult to get in the right order and proportions. Rotate it 90º to the left, then you can see the head out in front, the shell on the top / back, the feet hanging down underneath, the tail projecting behind, and the body in the middle of it all. So, even though it is troublesome to write, it’s easy to recognize and joke about. We soon learn that it is pronounced guī and means “turtle”.
Conversely, 茲 is relatively easy to write (9 or 10 strokes, depending on your preference), but a nuisance to remember what it means: “now, here; this; time, year”.
Put the two characters together and what do you get? Guīzī 龜茲 (“turtle this”).
Wrong, dead wrong. You have to give both characters an irregular reading and bleach them of their meaning. Instead of guīzī 龜茲 (“turtle this”), you must remember to read them as the transcriptional name Qiūcí 龜茲, which sinographically stands for Kucha.
Here's the etymology of Qiūcí 龜茲.
First attested in the Book of Han, written in the 2nd century. Apparently corresponds to the endonym of the historical Kucha in Tocharian B *Kuśi, attested in the genitive Kuśiñ and adjective kuśiññe (“pertaining to *Kuśi”), or its more archaic variant (see Adams, 2013, which reconstructs the Proto-Tocharian *kući(ye) proto-form for the adjective). Cf. Niya Prakrit (kucirajaṃmi, “kingdom of Kuci”). Adams (ibid.) suggests further etymology from Proto-Indo-European *ḱewk- (“shining, white”).
Wiktionary (accessed December 28, 2024)
By the way, 龜 has another irregular pronunciation, viz. jūn, used only to signify chapped or cracked skin. I think it must reflect the same morpheme as jūn 皸 ("chapped; cracked [skin]").
May you have no turtle skin throughout the coming New Year!
Selected readings
- "An early fourth century AD historical puzzle involving a Caucasian people in North China" (1/25/19)
- "Tocharian C: its discovery and implications" (4/2/19)
- "A Kuchean shift in terminology from Indo-Iranian to Tocharian" 6/10/24) — with extensive classified bibliography
- "The origins and affinities of Tocharian" (8/20/23)
- "The geographical, archeological, genetic, and linguistic origins of Tocharian" (7/14/20)
33ə said,
January 1, 2025 @ 3:40 pm
Another irregular pronunciation to test your knowledge in this field of ancient geography in central Asia is 吐谷渾.
Gokul Madhavan said,
January 2, 2025 @ 5:45 am
I am fascinated by the pinyin version of Kumārajīva’s name (from the Wilkipedia link above): Jiūmóluóshí. The first three syllables seem to correspond to Kumāra, but if I hadn’t known the connection there is no way I would have made the association myself. I’m guessing the initial jiū comes from the palatalization of an older ku (similar or perhaps identical to the shift in Peking > Beijing). But why do the relatively open ā and a sounds in the original turn into o here? And of course, connecting back to the etymology of ramen, when did the r in his name become an l? That other post seemed to suggest the conflation of the two sounds is primarily a Southern Chinese phenomenon, but here it appears to have happened in Mandarin.
JimG said,
January 2, 2025 @ 8:22 am
… The oldest Chinese name I know of for Kucha is Qiūcí 龜茲…
In my ignorance and lack of knowledge, I wondered briefly if the Chinese name for Kucha might hark back to origins of the concept of a World Turtle, and the modern phrase "Turtles all the way down."
David Marjanović said,
January 2, 2025 @ 8:50 am
This, too, is a regular sound change. It is very common the world over; Bengālī pronounces itself with a rounded [ɒ], English has done the same thing (words like stone, home, foam, oak had ā 1000 years ago), and I could probably come up with 10 more examples from Europe alone.
It was a [l] to begin with, as the OP and several comments say, because the word comes from China and only made it to Japan rather recently.
Victor Mair said,
January 2, 2025 @ 8:57 am
@JimG
I absolutely love your comment.
Thanks!
Lucas Christopoulos said,
January 2, 2025 @ 9:17 am
Does Kucha city mean "Turtle" when they founded it? Kaccap means Turtle in Tocharian, which seems close to Kucha and the Chinese Qiūcí (龜茲).
Probably not connected, in ancient Greek turtle is khelônê, and the city of Argos (Argl) also comes from the etymology "white, silver, shining."
kaccāp (n.[m.sg.]) ‘turtle, tortoise’
[kaccāp, -, -//-, -, kaccāpäṃ] wlawātai anaiśai kaccāp ram no ṣañ lyñā/// ‘thou hast behaved carefully like a tortoise in his own shell’ (243b4), snai prenke takoy sa kenä … wars=ite eśnesa menkitse tākoy kacāp ompä pärkre-śāyeñca ‘the earth must have been without island and full of water; the tortoise there must have lacked eyes [but been] long-living’ (407a5/6). ∎From BHS kacchapa-
Gokul Madhavan said,
January 2, 2025 @ 11:14 am
@David Marjanović:
Thank you for your response. I should have been clearer on my real underlying question: how much of the (to my ears, substantial) phonetic gap between the Indo-Tocharian name kumārajīva and the modern Mandarin jiūmóluóshí can be explained by the phonology of the original Sinitic language into which it was first transcribed, and how much of it is attributable to regular sound changes between that initial Sinitic transcription and modern Mandarin? Thus, for instance, it seems like we are in agreement that [ku] > [tɕjy] is largely accounted for a sound change in Mandarin. If I understand your comment correctly, the shift of [ɘ] and [ɑː] to [o] was also similarly an intra-Sinitic development. However, per my reading of your comment, it would seem that the Sinitic speakers who first met Kumārajīva already did not distinguish between [r] and [l] and so would have pronounced his name with the [l] from the very beginning itself. Is that right?
Looking through Wiktionary, it appears that a Japanese reading of the characters 鳩摩羅什 could be ku-ma-ra-jū, which is somewhat closer to the original name, though partly due to the Japanese flip of the [l] to the [r]. (Disclaimer: I’m not claiming the Japanese pronounce Kumārajīva’s name this way today, or that they did so thus when first introduced to his name! I don’t know any of those things, but am curious to find out.)
If we were to do a similar exercise with Wiktionary’s Middle Chinese reconstructed pronunciations for these syllables, would that be our most accurate current reconstruction for how a Sinitic contemporary of Kumārajīva would have heard his name?
Gokul Madhavan said,
January 2, 2025 @ 11:32 am
@Lucas Christopoulos:
That is a very intriguing suggestion. Although I don’t know any Tocharian and cannot comment on the plausibility of your suggestion, I am struck by the fact that the Sanskrit word kacchapa / kaśyapa does not have a secure etymology. Alexander Lubotsky includes it in his “Indo-Iranian substratum” [1] as a word that was likely borrowed in the Proto-Indo-Iranian stage of the language. If it was the BMAC / Oxus civilization that was the source of this word, then could it also have directly loaned the word to the Tocharians (assuming the timelines work)?
[1] Lubotsky, Alexander. “The Indo-Iranian substratum.” Originally appeared in: Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. Papers presented at an international symposium held at the Tvärminne Research Station of the University of Helsinki 8-10 January 1999. (Mémoires de la Société Finno-ougrienne 242.) Chr. Carpelan, A. Parpola, P. Koskikallio (eds.). Helsinki 2001, 301-317.
Jonathan Smith said,
January 2, 2025 @ 1:39 pm
@Gokul Madhavan
re: Kumārajīva, using Pulleyblank's "Early Middle Chinese" from his 1991 lexicon you get kuw-ma-la-dʑip, or using Schuessler's "Later Han Chinese" (2009) you get equivalent ku-mɑ-lɑ-dźip. I think. Looking at other "reconstructions" here, differences appear small, but Pulleyblank did make important adjustments to certain syllable types (note /ku(w)/ not /kiu/ or sth.) which are adapted by Schuessler. Right, Chinese from Han or whenever forward is supposed to have already been affected by earlier r- > l-. Other funky stuff could still be going on of course, e.g. conservative spellings…
Yves Rehbein said,
January 2, 2025 @ 4:03 pm
The Bactria Margiana Archaeological Complex (BMAC) is separated by more than a couple hundred years from Buddhist Hybrid Sanskrit (BHS) and the same holds if Proto-Indo-Iranian should have borrowed from BMAC before BHS could be transmitted to Kucha. I believe kaccap would look a lot different from kacchapa were it so old.
I am not sure of the parameters for the comparative method in this respect. The situation of Proto-Tocharian relative to Proto-Indo-Iranian speakers after 3500 BCE, in the selective reading ("The origins and affinities of Tocharian"), is many more years older, still. For analogy, the Mittani-Indo-Aryan loanwords are safely attested in the second millenium BCE far to the west of the Hindus valley, south of the Caucasus, because writing has devoloped early in this part. Qiūcí 龜茲 or the tortoise words for that matter appear much later. The BMAC substrate is not affiliated with any language family.
"No etymology" is probably an understatement if dedicated literature is liable to entertain educated guesses. German Schildkröte "turtle" is somehow akin to Kröte "frog" and metonymically to Kaulquappe "tadpole", one compares Proto-Slavic *žaba < *žěba "frog, toad". Compared to *zmь̀jь "snake, dragon", if derived from Proto-Indo-European *dʰéǵʰōm "earth" https://en.wiktionary.org/wiki/Reconstruction:Proto-Indo-European/dʰéǵʰōm a toponym would match very well and cognate human might work as well for the ethnonym, in genetive case: "earthling".
Yves Rehbein said,
January 2, 2025 @ 4:37 pm
Re: Kumārajīva and MC kuw-ma-la-dʑip, I can relate that Yves becomes Vietnames ip. But you can say pho, I ask. It's difficult.
Chris Button said,
January 2, 2025 @ 10:26 pm
@ JimG
Joking aside, I wonder if you might actually be onto something here.
guī 龜 "turtle" is commonly reconstructed with a medial -r- to give Early Middle Chinese (EMC) kwi from Old Chinese (OC) qrəɣ (via qʷrəɣ)
Qiū 龜 in Kucha is commonly reconstructed without a medial -r- to give EMC kuw from OC qəɣ (via qʷəɣ). However, note that the Mandarin reflex should be Jiū rather than Qiū (Pulleyblank has a comment on this in his Lexicon).
Since OC -ʷəɣ gives EMC -uw, while OC -əɣ on its own gives EMC -ɨ, it occurs to me that guī "turtle" perhaps represents a variant pronunciation where the labial feature failed to spread rather than a medial -r- having blocked the spreading.
This would parallel the other OC syllable type with weight on the coda rather than the nucleus (OC -ᵊɣ giving EMC -əj) where the labial feature fails to spread: OC -ʷᵊɣ gives EMC -əj)
The failure of the labialization to spread could perhaps be caused by the loan origin of the word (discussed in Schuessler's etymological dictionary) leaving it to vacillate between the two OC syllable types.
Chris Button said,
January 2, 2025 @ 10:28 pm
"OC -ʷᵊɣ gives EMC -əj" should say "OC -ʷᵊɣ gives EMC –wəj"
Chris Button said,
January 2, 2025 @ 11:15 pm
To be clear, I'm not suggesting that Kucha meant "turtle" but just speculating that guī 龜 "turtle" and Qiū 龜 Ku(cha) might jisr perhaps be reconstructed in the same way.
AntC said,
January 2, 2025 @ 11:59 pm
Beware them ancient turtles: "large carnivorous turtle". I noticed an uncommon sense in that report:
Process 6. (anatomy) An outgrowth of tissue arising above a surface, such as might form part of a joint or the attachment point for a muscle. [wikti]
Rodger C said,
January 3, 2025 @ 10:56 am
Next we check whether the jū in ku-ma-ra-jū was written ji-fu in pre-1945 furigana.
Chris Button said,
January 3, 2025 @ 12:16 pm
什: Go-on じゅう jū, earlier written じふ jifu, which goes back to ⁿzip with a final u then emerging to bring Sino-Japanese in line with the phonotactic constraints of Japanese as ⁿzipᵘ.