*g-wan

This etymon is represented in the Naga, Qiangic, and Himalayish branches. The Amdo (Bla-brang) form hwoŋ wa seems to represent a distinct root. The velar prefix is attested in both Ntenyi and Lepcha.

The second syllable of Ersu da⁵⁵ vɛ⁵⁵ ȵo³³ looks like a possible reflex of this etymon, but vɛ⁵⁵ȵo⁵⁵ is glossed as ‘intestines’. However, the semantic development of HAND + INTESTINES > ARM seems quite implausible.

There is a good Chinese comparandum meaning ‘wrist’.

腕 OC *ʔwɑ̂n, GSR #260n ‘wrist’; Li 1971: (*·uanh); Baxter 1992: (*ʔons); B & S 2011: *[ʔ]ˤo[r]-s; Schuessler 2007:582-3 (see 迂 below); Mand. wàn.

捥 OC *ʔwɑ̂n, GSR #260m ‘wrist’; Li 1971: (*·uanh); Baxter 1992: (*ʔons); B & S 2011: *[ʔ]ˤo[r]-s; Schuessler 2007:--; Mand. wàn.

ZH: The Old Chinese words may have ended in *-n or *-r, which merge in later stages of Chinese; the brackets around *r in Baxter & Sagart’s (2011) reconstruction are meant to reflect these two possibilities. Bodman (1980:65) suggested a comparison with Old Tibetan ’khord ‘to turn around’ > WT ’khor and reconstructed OC *skwars.

Schuessler 2007:582-3 instead derives the Chinese word for ‘wrist’ from an open-syllable root meaning ‘bend’ (迂, Mand. yū), with nominalizing suffix *-n, an etymology unrelated to this PTB root.

OC forms reconstructed with *o (*wa in Karlgren and Li) should normally correspond to PTB *o, not *wa. But the support for this root in TB is sparse enough that a reconstruction with PTB *o probably can’t be ruled out; besides, comparisons involving zero or glottal initial may have different correspondences that remain to be clearly worked out.