Comments:ТМС 2, 105, Doerfer MT 145. Cf. also Kalm. soɣǝ 'a k. of big plant; a k. of reed' (KW 329). In the Kor.-Jpn. area there may be some confusion between the reflexes of this root and PA *sokŋV 'onion': Kor. sɨŋ'a and modern Jpn. forms like sukampo and suiba may actually go back to the latter.
Comments:Medial *-j- should be reconstructed to explain loss of *-l- in Kor. The root seems to be a Wanderwort (see Аб. 3, 194-195, Cтеблин-Каменский 1972, 31, NCED 965), but may be reconstructable for PA.
Comments:The comparison with Jpn. sakana 'fish' in JOAL 98 should certainly be rejected (the only meaning of sakana attested in OJ is "food (vegetables) eaten with sake", thus the modern meaning 'fish' is obviously secondary). Cf. *sàk`ò.
Comments:The Turkic and Mongolian forms basically denote 'tamarisk' or 'tamarisk branches', but whether this meaning was original is not quite clear. Perhaps the root just denoted dry branches (used basically as firewood).
Comments:Martin 233. Jpn. *sua presupposes a suffixed form *súm(e)-gV. The root can be traced in all branches, but everywhere is rather scantily represented and thus not quite reliable; the exact sort of the denoted plant is also not quite clear.
Comments:KW 337, Poppe 30, 70. A Western isogloss, and a quite exceptional case of a V̄ : V̄ correspondence between PT and PTM (expressive root? or an old loanword, cf. Doerfer MT 46 ?). On the possibility of Mong. < Turk. see Щербак 1997, 147.
Comments:Cf. also Evk. sirahun 'whey' (possibly reflecting an unattested Mong. *sira-sun). MMong. širxe 'vinegar' (HY) is most probably borrowed < Turk., see Щербак 1997, 145. The vowel in Turkic is not quite clear (*sürke or *sörke would be expected); cf. also Karakh. (MK) sorquč (perhaps sarqɨč or sɨrquč - vocalization is uncertain) 'a k. of spicy plant (chicory?)', see EDT 848.