Comments:SKE 83, Lee 1958, 111, МССНЯ 356, ОСНЯ 1, 225. One of the few PA monosyllabic roots. The relation of the Turk. form is somewhat dubious (for semantic and phonetic reasons). One may conjecture that the original shape was PT *Ka- - actually attested in OT, see EDT 578 - which was early substituted by *Ko- under the influence of the more widespread *Kod- and then disappeared altogether.
Comments:Rozycki 85 (TM-Mong.). Despite Doerfer MT 51, TM cannot be borrowed from Mong. The Mong. form presents some difficulties: apart from the unclear variant *ganča, the form *gagča itself can only belong here if it is an irregular development < *gagči < *gagti, or if it is a contraction of a derived form *gagta-ča. Note that Jpn. *kàtà may also be derived from PA *kalt`o q. v.
Comments:Мудрак Дисс. 102. Cf. also Kalm. gejǖn, WMong. gejeɣün 'grievous, gloomy' (perhaps < *gaju-gi- with secondary fronting). Note the peculiar *-d- in PT, reflected as -j- (not -r-) in Chuv.: perhaps we should rather reconstruct *Kaj(i)-dgu for early PT, with subsequent development either > *Kajgu or > *Kadgu.
Comments:Ramstedt (SKE 415) thinks of a loan Kor. > Turk., which is highly improbable (see TMN 1, 415); despite late attestation in Turkic the root may well be archaic. For TM one should suppose the original meaning "children of smb.'s concubine(s)".
Comments:Martin 251 (Kor.-Jpn.), Lee 1958, 111 (Kor.-TM). Fronting in Turk. is unclear (*Kaŕ would be normally expected). The Kor. word may also reflect *kằle q. v. (Joki 1963, 154); in any case, it actually reflects a suffixed form *gàŕ[a]-KV (cf. Man. GarGa) or *kale-kV.
Comments:KW 145, Владимирцов 391, Poppe 24 (Turk.-Mong.; but ТМ *ŋāla 'hand, arm' cannot belong here), Колесникова 1972а, 95-97; АПиПЯЯ 290, Дыбо 310-311, Лексика 247. Borrowing in Mong. < Turk. is quite improbable, despite Щербак 1997, 134 (even Doerfer in TMN 1, 207, 3, 461 describes the Turk.-Mong. match as "Zufall" - which in all his works is actually a synonym for "cognate").
Comments:A Mong.-Tung. isogloss. On a possible reflex in PT see under *gā̀ŕV; if PT *Kāŕ at least partially reflects the present root, it would be an argument to reconstruct PA *gāso with a long *ā. One should also note Kor. dial. kesani 'goose' (KED 109) - if it is not a derivative of PK *kǝ̀jú 'goose' (see under *kúja).
Comments:KW 152, SKE 128, Poppe 24, Ozawa 89-90, Miller 1986a, 55. The development 'fur on paws' > 'shoe' in Mong. and Jpn. may well have been independent, but one has also to reckon with a possibility of ancient borrowings (especially because of some phonetic irregularities in Mong.).
Comments:EAS 113, Poppe 18, Цинциус 1972, 6, АПиПЯЯ 80, Лексика 171, TMN 3, 387 ("onomatopoetisch"). The Turk. form is probably contaminated with *Kās == TM *gasa 'crane, duck', which should explain the exceptional preservation of vowel length. See also comments to *ki̯ujilV and *gi̯ằla.