Comments:EAS 73, VEWT 169, SKE 62, 59, АПиПЯЯ 291. Korean has ǝ- instead of the expected *ă, because the latter cannot stand in Anlaut. A derivative of the same root may be MKor. ón 'hundred' ( < *ēgo-nV). {Cf. PE *uɣla- 'numerous, many'.}
Comments:Ozawa 8, Дыбо 12. Medial *-j- has to be reconstructed to account for lack of spirantization in Mong. and for -w- (not -p-) in Jpn. Note a similar dental suffixation in Mong. and Jpn.
Comments:EAS 145, KW 118, Poppe 76, ОСНЯ 1, 268, Rozycki 68. A Western isogloss (in fact, basically Turk.-Tung., since Mong. may be < Turk., see TMN 2, 200; Clark 1980, 43 on Mong. elči < Turk. ēl-či).
Comments:VEWT 43, Martin 234. Note a common Altaic derivation in *-kV, reflected in all languages except Turkic; for TM we suppose a semantic development 'sheath' < 'receptacle' < 'inside, middle'. Jpn. has lost the first vowel in this archaic trisyllabic structure, and Kor. has rising tone also due to contraction.
Comments:ЭСТЯ 1,154 (Turk.-Mong.), АПиПЯЯ 286. The Jpn. word poses some problems: its accent does not correspond to Turkic length, and if we reconstruct *ēŋo (as required by Jpn. *ǝ-), PT should have a closed *ạ̄. It is worth while therefore to consider other explanations of Jpn. *ǝ̀mǝ̀-p-: e. g. compare it with Mong. ojun 'thought' and reconstruct a separate protoform like *uńo.
Comments:KW 19, Poppe 46, 97. Cf. Manchu ebte ǵaxun 'hunting bird' (perhaps preserving the original shape of the root *eb(a)-, that has elsewhere been irregularly reduced to *ewā- > *wā-). Despite TMN 1, 107, 2, 131, Щербак 1997, 94, there is no reason for assuming Turk. > Mong.
Comments:EAS 107, 145-146, SKE 56 (Mong.:Tung.:Kor.). Mong. and Turk. have also suffixed forms that may belong to the same root (with front / back vowel variation): cf. Mong. arča- 'to quarrel, argue'; erüs- 'to rival, compete'; Turk. *ersi- 'mean, nasty' (see Дыбо 12); perhaps also Mong. *arila- 'to disappear, vanish' ( < '*to miss'), see MGCD 118.
Comments:EAS 146, KW 126, VEWT 46, SKE 69, Poppe 106, Menges 1984, 267, Дыбо 11, Лексика 79. Cf. perhaps also MKor. jǝ̄t 'now' ( < *jǝ̄r-t = Mong., Turk. *erte?); Lee 1958, 108 also draws MKor. ǝri 'time, season', ǝrɨn 'quickly', which we were unable to identify. Borrowing in Mong. from Turk. (Щербак 1997, 116) is hardly plausible: the *-t`V suffix in this case must be Common Altaic. The etymology is still quite valid, despite all attempts of Doerfer (TMN 4, 257-259). Cf. also Turk. *er-k- 'early; ancient' = Mong. argi- 'to be old, ancient, mature'.
