Comments:ОСНЯ 2, 86, АПиПЯЯ 295. See SKE 159, EAS 75 (Mong.-Kor.); one frequently links Turk. *jāŕ 'spring' (see SKE ibid., KW 272, VEWT 193), but the latter should be separated (see *ni̯ā́ŕ[à]); instead it seems plausible to compare Turk. *jarɨ-n 'tomorrow, morning' - see Лексика 80-81. TM *ŋēr(i)- 'light' is a perfect phonetic and semantic match and should be separated from Mong. gere-l 'light' (especially because the rule *ŋ- > Mong. g- is most probably false) - despite KW 134, Poppe 25, ОСНЯ 1, 228-229, АПиПЯЯ 18, Дыбо 11; on the etymology of the latter see under *gari (despite Poppe 1972, 101, Doerfer MT 21, the TM form of course cannot be borrowed from Mong. gerel).
Comments:ТМС 1, 352. It is tempting to compare also Evk. ńewte, Evn. ńewte 'spring, well' (*'washing or pouring place') and perhaps also OJ mjiwo 'water-way, seaway' (if mji- is to be analysed as 'water', the -wo part stays completely obscure).
Comments:The reconstruction of the diphthong in this root is based on PT *j-: one should suppose an early development *ŋi̯ō- > *jo- in Turkic (which explains the vocalic reflex). Cf. *i̯úgu 'up, above' (which could also influence the Turkic form due to mutual influence of synonymous *jüg- and *jok-).
Comments:АПиПЯЯ 297. Turk. > Mong. al (KW 6, Щербак 1997, 97). The Jpn. parallel is quite convincing, despite a mismatch in tone (probably distorted in a long compound).
Comments:? Cf. also Turk. *üč (*öč) 'three' (ЭСТЯ 1, 641-642, Stachowski 254). The match is somewhat problematic, basically because of the absence of the TM cognate, shortness of the root and unclear suffixation. The basic phonological correspondence pointing to PA *ŋ- is, however, observed, and the parallel seems to be worth noting.
Comments:АПиПЯЯ 18, 55, 81, 104, 278 (with literature). Initial *n- in Jpn. is probably due to assimilation to the following nasal (the most usual form is *na-ni); there also exists an interrogative *i- (in *i-ka 'how', *i-n-ture 'which' etc., reflecting nasalless *ŋi-. Vocalism in this archaic monosyllabic pronoun is not quite clear, evidently because of different suffixation.
