Comment: Cf. also Tab. Düb. šurš (pointing to a week *ʎ). The meaning 'moss' in Lezg. is clearly derived from 'slime', and the meaning 'nose'in Rut. - from 'snot'. Obl. base *ʎerʎe- (cf. Ag. Burk. xirxi-, Rut. xexi-). 3d class in Rut., but 4th class in Tsakh. The Tab. form (or some unattested Ag. dialectal form) was borrowed in Darg. Chir. as šerš-ne (plur.) "spittle".
Comment: The root is isolated in Tsakh. (cf. also Gelm. xanä; its precise meaning is 'bottom (as seen from below)', as opposed to ḳan 'bottom (as seen from above)'), with a probable Lak. parallel.
Comment: Obl. base *ʎ:ilV-, cf. Rut. xilɨ-, Ag. xilu-, Lezg. (with vowel levelling) gele-. Cf. also (for the reflexes of *ʎ:-) Tab. Khür. š:il, Ag. Bursh. š:il, Tsirkh. x:il. 3d class in Rut. and Kryz. The Tsakh. form probably reflects the original meaning "foot".
Comment: Other relevant forms: Tab. Khür. š:av (pointing to PL *ɫ:), Düb. šav. Traces of the PL oblique base *ʎ:im:V- (with Ablaut) are preserved in Rut. xäb,xibɨ-l-; the Tsakh. form (cf. also Gelm. xɨbɨna) reflects the original plural *ɫ:im:V-p:V with metathesis. Metathesized is also the Ud. form muχ - but external comparison shows that metathesis must have occurred in the rest of PL dialects after the separation of Udi.
Comment: The Lezg. word is preserved only in compound kinship terms: iran-buba 'father-in-law' (*"wife's father"), iran-dide 'mother-in-law' (*"wife's mother"), iran-stχa 'brother-in-law' (*"wife's brother"), iran-waχ 'sister-in-law' (*"wife's sister"), with dialectal (Yark.) variants jaran-. The oblique base can be, therefore, reconstructed as *ʎ:ɨra- or *ʎ:ɨrä-. (One should note that the Ag. -r- here is present in all dialects, cf. Fit., Burk. xir, and thus can not go back to *-n:-; it also corresponds quite regularly to Tsakh. -w-. Thus this root must be kept apart from PL *ʎ:ɨn:- q.v.).
Comment: Cf. also Lezg. Kur. jiṭ 'boiled wheat'. 3d class in Rut., Tsakh. and Bud. Obl. base *ʎ:ɨwṭä- (besides Tsakh. xɨṭa- cf. also Rut. xɨṭa-). Labialisation (either -wṭ- or -ṭʷ-) must be reconstructed to account for the labialized Bud. reflex f- (fiṭ < *xʷiṭ < *xiwṭ); however, there is no way to recopstruct *ʎ:ʷ- in this root, because of uniformly non-labialized reflexes in all other languages.
Comment: Initial ʕa- in Ag. is unclear (cf. in other dialects: Bursh. jiš̌u-r, Fit. jüfü-d, Burk. ifa-d; here the PL numeral obtained the initial ji- (je-) on analogy with *jewq̇ɨ- 'four' q.v.).
Rather interesting is the Archi word for 'fifty': buχ:í (not buʎ:í). Since Archi names for tens are usually formed from simple numerals, this form presupposes *χ:ʷe 'five' - which can reflect PL *χ:ʷä (from a lax PEC variant *fɦä̆).
