Comments:EAS 65, KW 479, Владимирцов 197, SKE 31, Poppe 28, 111, Дыбо 13. The comparison seems quite valid, despite doubts in TMN 4, 225. Despite Poppe 1966, 191, Doerfer MT 24, Rozycki 128, TM is hardly < Mong. Loss of *-g- in Turkic is probably explained by the fact that the root is only used with suffixes (*jüg(V)k > *jük, *jüg(V)d > *jüd-).
Comments:АПиПЯЯ 77, Rozycki 124. Tungus forms may be borrowed from Mong. (see Poppe 1966, 194, Doerfer MT 80). Medial *-j- has to be reconstructed, as in a number of other cases, to explain Mong. -b- (not -ɣ-) and Jpn. *-w- (not -p-); exceptional back vowel in Turkic may be also due to dissimilation with the medial *-j-. In Turkic one has to suppose a semantic development 'thin' < 'lean, weak' (the latter meaning is actually widely attested since the earliest texts).
Comments:SKE 38; Poppe 1950, 572, Lee 1958, 113, Menges 1984, 266, АПиПЯЯ 77; Дыбо 5, Martin 231, Miller 1985, 150, 1985a, 83, JOAL 99, Whitman 1985, 126, Лексика 276. The root is quite reliable, but not completely regular, probably due to contractions and assimilations (not quite clear is -i- in Mong. - on analogy with ǯirüke 'heart'? - and voicing of -k- in TM).
Comments:KW 483-484, Владимирцов 188, АПиПЯЯ 282. One cannot exclude Mong. being borrowed from Turkic (see Щербак 1997, 126), but there are no special reasons for that.
Comments:The etymology is worth listing, but not devoid of problems: in Mongolian we have only an isolated Kalmuck form, and the Tung.-Jpn. equation is possible only if the variant *ǯōn- in TM is secondary.
