Comments:SKE 138-139, PKE 103, Menges 1984, 277. Despite Doerfer MT 144, hardly borrowed in TM < Mong. In Turk. the root has probably merged with *boŕ- / *buŕ- 'damage, oppress' (cf. also *burul- ~ *borul- 'to be angry', ЭСТЯ 2, 269). The Korean form may be alternatively compared with Mong. melǯe- 'to refuse, refrain'.
Comments:Martin 229, Whitman 1985, 127, 194, 239, SKE 153, АПиПЯЯ 279. An Eastern isogloss. In MKor. cf. also mɨrɨs 'all, in general'. In Jpn. low tone would be expected; high tone here was probably induced by the similar *mǝ́r- 'to heap up' (see under *mi̯ólo), which also influenced the vowel in *mǝ́rǝ́.
Comments:SKE 151 (Turk.-Kor.), АПиПЯЯ 288. In MKor. cf. also a reduplicated form: màmằrằ- 'rough, stony (ground)'. The Jpn. form must be explained as a result of nasal assimilation (*mana- < *mara-n-); but cf. for Jpn. *mana-n-kua alternatively Mong. *maŋka 'long sandy hill' (MGCD 481).
Comments:EAS 109, SKE 146, АПиПЯЯ 31-32, 282, Мудрак Дисс. 90, Лексика 194. Doerfer's (TMN 2, 253) criticism is short ("unklar"). On possible traces in Jpn. see under *k`ŏjli. An unsuccessful attempt of refuting the etymology was undertaken by Vovin 2000, who argues that the attested Old Korean form is 麻帝 MC mạ-tìej [ma-te]. However, it is most probable that MC -t- was used here just to transcribe Korean -r- (since Middle Chinese, as well known, lacked r-). Anyway, it is hardly possible to make any decisions on the basis of very inadequate and scanty Kirim transcriptions.
Comments:The Kor.-Jpn. form is a derivative in *-rV (*-lV). The Tungus form fits very well semantically, but raises some phonetic doubts: vowel length and the quality of -n- (instead *-ń-) do not correspond to other languages. Cf. perhaps Turk. (Oyr.) man 'fence' (VEWT 325). See Vovin 1993, 257.