Did labialized consonants remain in Ryūkyūan?

Question

The labialized velar consonants are rare even in historical Japanese; besides (possibly) one verb (kweru), they (kwa-, gwa-, kwe-, gwe-) occur only in some Sino-Japanese morphemes. In Japanese, they merged with the non-labialized versions (ka-, ga-, ke-, ge-) as late as the early 20th Century.

But do they remain in Ryūkyūan?

Or, did they even exist in Ryūkyūan? Obviously, Ryūkyūan also borrowed multiple Sino-Xenic terms, some of which must have been necessarily present: such as name of the goddess Kwannon. Do we know the current status of words like this? Even if such words are Sino-Japanese borrowed from Japanese, did the labiovelars remain in Ryūkyūan even when, very recently, Japanese stopped pronouncing them with the //-w-// glide? Unfortunately, the vast corpus of scholarly work, to my knowledge, rarely deals with the Sinitic layer in Ryūkyūan.

Answer 1

After poking around a bit in the Shuri-Naha Dialect Dictionary that covers the Okinawan branch of Ryūkyūan, it appears that here, at least, labiovelars did persist. A sampling of entries that show maintained labiovelars:

• クヮンジミ, Japanese 缶詰【かんづめ】
• クェーチウイウェー, Japanese 快気【かいき】祝【いわ】い
• クヮンヌン, Japanese 観音【かんのん】
• カングヮーシ, apparently equivalent to Japanese 羊羹【ようかん】 + 菓子【かし】; the entry also mentions ムイグヮーシ with a given kanji spelling of 盛菓子.
• グェーシチ, Japanese 外戚【がいせき】
• クェーテー, Japanese 懐胎【かいたい】

I did find a term where Japanese dictionaries indicate an older //gwa-// pronunciation, but Okinawan doesn't reflect the labiovelar.

• ガンジュー, Japanese 頑丈【がんじょう】

This appears to be an exception to the rule, however, which raises the question of whether this was a later borrowing from Japanese, rather than an older import when the kanji term was borrowed from Middle Chinese.

Notably, the labiovelars I can find reflected in Okinawan are all traceable to older //kwa-// or //gwa-//. The //kwe-// and //gwe-// variants found in Japanese appear to have undergone different sound shifts in Okinawan, such that the labiovelar was apparently lost. One clear example I could find was トゥチー, cognate with Japanese 時計【とけい】. The likely original, Chinese-derived spelling was 土圭, and 圭 apparently had an older reading of //kwei//. So it seems that the //kwe-// labiovelars may have shifted first to something like //ki-//, and then underwent the regular change to become affricate //t͡ʃi-// (incidentally, much like the pronunciation of ⟨ci⟩ went from Latin //ki// to Italian //t͡ʃi//).

---

Please comment if the above does not fully address your question.