Comments:SKE 237, EAS 71, Martin 245. Doerfer (TMN 3, 252) protests against comparing Turk. and Tung. because TM *u (Benzing's *ö) does not correspond to PT *ü. This is of course not true; but the root still has a slight irregularity, namely, high tone in Korean (generally not typical for a verbal root, and not corresponding to Turkic length).
Comments:АПиПЯЯ 277. The root is rare in Altaic, but has reliable external parallels (see МССНЯ, 1, 7). It may also be represented in Turk. *sajɨ- 'each, every', Mong. sajiki 'the same').
Comments:Phonetically OJ subaru is a quite satisfactory match for Mong. and TM forms, so Miller's (Miller 1988) attempts to derive it either from Arabic al-zubra 'mane' (name for the two stars of the 11th lunar station) or from MKor. spɨr 'horn' are probably not necessary.
Comments:Владимирцов 248, Цинциус 1984, 12-13, АПиПЯЯ 293. A Western isogloss. The Mong. variants *čaji- / *čeji- reflect most probably a merger with the root *šā̀ŋu 'clear, light' q. v.
Comments:Mong. has also *čaji- / *čeji- 'be white, whitish', associated in the modern language rather with *čaga-ɣan 'white' (see *šā̀k`a), but phonetically rather going back to *čaŋi- < *šaŋu. Despite Rozycki 193, the TM forms are hardly borrowed < Mong.
Comments:KW 423, Poppe 63 (Turk.-Mong.). Mong. is not < Turk., despite TMN 3, 217, Щербак 1997, 144. There are some phonetic uncertainties: Korean has also a variant čǝ̀čí- 'to soak' (see PKE 30); in TM one observes a variation between *-i- and *-e-.
Comments:A common derivative *šèru-p`V- is reconstructable for Turko-Mongolian (PT *sara-p, *sara-pa-n, PM *sara-b-či). The root evidently denoted some sort of lattice used in construction.
Comments:EAS 150, Владимирцов 212, Poppe 30, 47, Мудрак Дисс. 43-44. Mong. cannot be explained as a loanword, despite Щербак 1997, 144, or as "bloßer Zufall", despite TMN 4, 288. In Kor. one has to suppose a secondary development *psɨ́r- < *spɨ́r-.
Comments:АПиПЯЯ 277. Initial *š- must be reconstructed because of Kor. *s-. Kor. sà'òrí appears to be genuine (cf. the match with Jpn. *súwá-r-), but a loan from Mong. saɣuri(n) 'seat' (whence certainly Manchu sōrin) cannot be excluded, see Lee 1964, 192-193.
Comments:The root is rather sparsely attested in Turk. and TM, and the vocalic reconstruction is not quite secure (note that Jpn. -mp- speaks rather against the reconstruction of diphthong - unless the form is not actually a contraction of *súwu-mp- with a labial suffix).
Comments:The TM form is somewhat irregular: we would rather expect *šiči. Note, however that in Manchu we may be dealing with assimilation (čeče < *šeče), while other TM forms may in fact be borrowed from Manchu.
Comments:Дыбо 311 (Mong.-Tung.); Martin 232, Miller 1986a, 49, АПиПЯЯ 95, 279 (differently). The root must be kept distinct from *si̯ŏlo q. v., but contaminations were possible.