Comments:Цинциус 1984, 92, Ozawa 80-81. The stem may be connected to PT *kert- (ЭСТЯ 5, 54, Stachowski 145), Mong. *kerči- 'to cut' ( > Evk. kerči- etc., ТМС 1, 453, see Doerfer MT 110; the two words were compared in KW 228, Poppe 19, 51, 83, Menges 1982, and despite Щербак 1997, 127, Mong. kerči- is hardly borrowed < Turk.): we may be dealing with two derivatives (*k`ire-gV vs. *k`ire-t`V). However, direct derivation PT *kerki < *kert-ki is hardly plausible, despite Menges 1944; it rather goes back to *kirge-ki.