Comment: In Agul dialects there is a peculiar reflex of PL *-rƛ:-, cf. Rich. adi-s, (but after the preverb q-: q-aji-s 'to return'), Bursh. argi-s, Burk. are-s, Tp. ade-s. The Ud. form can be compared if we assume a reduction of the initial vowel. In Tsakh. cf. also s-ale= 'to gather'.
Comment: The Kryz. and Bud. form reflect the PL Ablaut grade *ʔirq̇I:ä-; the Auslaut *ä is reconstructed on basis of Ag. (Bursh. raqI:a-s) and Rut. (g-aq:a-s, pl. ga-q:e-s). Not quite clear is the loss of pharyngealization in Rut. See Хайдаков 1973, 121.
Comment: An isolated Tsakh. word (with possible Tsezian and Andian parallels). The synonymy "face": "sour cream" is explained by a usage present, e.g., in Udi, where we have muč:Ianaqin č:Io "sour cream", lit. "milk's face".
Comment: All forms point to weak *-s- in PL (cf. also Tab. Düb. ars), thus Ag. Bursh. ars: (MSU recording) must be a misspelling (cf. ars in Shaumyan's recordings). 4th class in Archi.
See Бокарев 1961, 60; Лексика 1971, 191; Гигинейшвили 1977, 68.
Comment: Cf. also Tab. Düb. k-a=š̌- (in Tab. the verb is attested only with preverbs), Arch. Dur. =ars:u-r. With Ablaut cf. perhaps also Tab. u=š̌-, Düb. u=š̌- 'to mow'.
An old derivate from this root is PL *[ħ]ars:ʷV-t:V (with an adjectival suffix, "cutting") > (with metathesis) > *[ħ]at:ers:ʷV "scissors", reflected in Tab. ubruš̌ (with assimilation), Düb. uruš:̌u, Ag. ħut:es (Bursh. ħaras:, Burk. ħäras), Tsakh. ades:e and Arch. jat:ǝrsi. We must note that this stem reveals some traces of contamination with another PEC root for "cut", *HVrsA (q.v.): the initial laryngeal and the irregular weak -s- in Archi must be traces of the other root, which otherwise vanished in Lezghian.
Meaning:1 to rub 2 to pick at, peck at 3 harrow 4 to comb
Lezghian:eχIä-d (Khl.) 3
Tabasaran:a=χI- 1,2
Agul:kit-X̀- 1
Rutul:a=χIa- 1
Kryz:ä=ħ- 1
Budukh:q:ɨ=ħ- 1,4
Comment: Cf. also Tab. Düb. ka-raχ- (with unclear loss of pharyngealization - or a misspelling?) 'to rub', Ag. l-arX̀- 'to thresh', q-arX̀- 'to touch', Bursh. k-erX̀a- id. (in kit-X̀- the root has the Ablaut grade *i: kit-X̀- < *kit-irχI-, cf. imper. kit-irX̀ etc.), Rut. ʁIa=χIa- 'to smear', Kryz. q:ɨ=ħ- 'to rub' ( = Bud. q:ɨ=ħ-), Bud. ja=ħ- 'to scratch, scrape'. In Agul the root is used with several preverbs and has acquired a wider meaning 'to touch, push > to drop, let fall' (e.g. al-arX̀- 'to fall, let fall' etc.).
Comment: j- and w- are prefixed class markers. In PL there existed an Ablaut grade *ʔorχ:V-, reflected in Kryz., Bud. and Rut. (where χɨrɨ- < *ʔorχ:ɨ- with a usual adjectival metathesis).
Meaning:1 to smear 2 to glue 3 to pull on 4 close 5 put on
Lezghian:hal-s- 3
Tabasaran:a=s- 1,k-a=s-2
Tsakhur:ǵ-a=s:an- 4
Kryz:q:-isn- (dur.)5
Archi:s:us:ǝ-bos 1
Comment: Cf. also Lezg. asun 'plastering, white-washing', Nüt. asã- 'to smear, plaster' (showing that the meaning 'to pull on' in Lezg. hal-s- is a secondary semantic development due to the preverb). Archi has preserved only a reduplicated stem (in a compound verb). The verb belongs to the strong PL conjugation class.
Meaning:1 to cut 2 to kill 3 to die (pl.) 4 to throw
Lezghian:aṭa- 1
Tabasaran:kt-a=ṭ- 1
Agul:aṭ- 1
Rutul:h-a=ṭʷa- 1
Tsakhur:ha=aṭa- 3
Kryz:ja=ṭr- 1
Budukh:o-ṭ- 2
Archi:a=ṭa- 1
Udi:bost:un 1,4
Comment: The verb belongs to the strong verbal class in PL. Initial Kryz. j- is probably a preverb (cf. in Bud. o-ṭ- 'kill': jo-ṭ- 'cut', ʕa-ṭ- 'pound, thresh', ʕo-ṭ- 'to push', e-ṭ- 'to cut in small pieces'; also Kryz. q:-o=ṭr-, Bud. q:u-ṭ- 'to hit', Kryz. q:-a=ṭr- 'to mow'). Cf. (for the second vowel) Lezg. praet. aṭu-na, Ag. Bursh. aṭi-s, Arch. dur. =arṭu-r. The Udi form bost:un goes back to *b-ot:-sun.
Comment: The Arch. form χ:ʷa means 'part of the mountain bordered with growth'. Cf. also Ag. Bursh. aχʷ 'wood, forest'. All the different meanings in Lezg. languages can be explained by the original meaning "mountain slope" > a) "woody place, wood, forest"; b) "slope > drain, ditch". The Arch. form probably reflects a former locative (in *-aʔ) with initial reduction.
Comment: Cf. also Tab. Düb. a=ʁI-, Ag. Bursh. ṭ-aX̀a-. The verb belongs to the weak PL class. The Ag., Rut. and Tsakh. forms contain an expressive preverb *ṭ-. In the Archi form b- is obviously a former class prefix; dur. babχ:Iu-r points to a secondary labialisation *-wχ:I- > *-bχ:Iʷ- in early Archi.
Comment: Labialisation in Lezg. is clearly secondary (all other languages point to *χ). The original meaning was probably 'to lie; to throw, fall', cf. Ud. b-arχi 'horizontal' ( < *'lying') (perhaps also, with prefixation and Ablaut, Rut. qa-ruχa-s 'to throw'). A local derivate from this root is Tab. aχin (Düb. aχin) 'bed', Ag. Bursh., Fit. aχun id., Rich., Burk. aχun 'mattress'.
Meaning:1 to know 2 to teach 3 be able 4 teaching,lesson
Tabasaran:aʁIu- 1
Agul:aħa- 1
Tsakhur:=aχI:a- 1,3
Kryz:ħa-r jeriǯ 2
Udi:aχI 4
Comment: A verbal stem, but mostly with a defective paradigm (only in compound gerund forms in Tab., Ag. and Kryz.); Udi has preserved an old deverbative noun.