Comment: -n- in Rut. (Khniukh) is due to contamination with another root, PL *q̇IangVr q.v.; external evidence shows that they are quite distinct, although they tend to merge in Lezghian languages.
Comment: Cf. also Lezg. Khl. q̇ük, Ag. Burk. ʡaj, Tp. ʡüj. Tab. -r-, as well as a special development of *-ƛ: in Lezg., point to a medial resonant in PL. This could only be *-n- (*-l- could not precede laterals, while *-rƛ:- yields other reflexes - Ag. -r(g)- and Lezg. -r-). Cf. a similar behaviour of *-n- in *menƛ:ä- 'eight' (with minor differences conditioned by different vocalism).
Comment: Both in Lezg. (lit.) and in Ag. the initial voiced consonant is a result of dissimilation ( < PL *q̇I-), cf. dialectal forms: Lezg. Khl. q̇üč̣, Ag. (Shaumyan) q̇ürč̣ 'moth'. The latter form also reveals a medial liquid which can only go back to PL *-l- (*-r- would have disappeared completely).
Comment: Cf. also Rut. Khnov q̇Iač̣ɨ-χI-dad 'armpit', Lezg. Khl. q̇üč̣ (qüč̣ü-) 'armpit', q̇Iʷäč̣ (q̇Iʷäč̣i-) 'nostril'; Tab. Düb. q̇Iuč̣u 'armpit' (in Kand. there occurred a dissimilation *q̇I-> ʁI-); Rut. Khnov. xexid q̇Iač̣ 'nostril' ("hole of nose"), Khniukh. q̇Iɨč̣ 'cave'.
A rather complicated case. For PL we have to reconstruct a paradigm with Ablaut *q̇Iʷač̣,*q̇Ioč̣ɨ- (preserved in Tab. ʁIuč̣i-); further levelling led to a division into two nouns in Lezg. (Khl. q̇üč̣ < *q̇Ioč̣,q̇Iʷäč̣ < *q̇Iʷač̣) and in Rut. dialects (Khnov q̇Iač̣ < *q̇Iʷač̣ vs. Khniukh q̇Iɨč̣ < *q̇Ioč̣).
In most Agul dialects (because of the change *q̇I > ʡ in some dialects) this root merged with PL *hIač̣ 'hollow, pit' (q.v.); only in the Burkikhan dialect there persists an opposition ʡač̣ 'armpit' (*q̇Iʷač̣) : ʕač̣ 'ravine'.
Comment: Cf. also Tab. Düb. q:Ianč, Ag. Bursh. q:Ianč:, Fit. q:Iʷanč:, Burk. q:ʷenč:. The final consonant is more or less regularly reflecting PL *č̣:, but the initial suffered assimilations (*q̇Iʷ > q:Iʷ) or dissimilations (*q̇Iʷ > qIʷ) almost in all dialects. Obl. base *q̇Iʷanč̣a- or *q̇Iʷanč̣ä- (cf. Ag. q:Ianč:a-, Rut. qIʷanč̣a-). 3d class in Rut.
Comment: 4th class in all class-distinguishing languages. The medial resonant is not clear: Bud. seems to point to *-l- or *-r- (assimilated to the final -l), but other languages do not reflect any liquid. The stem is probably derived: cf. Arch. q̇Ieṭ 'a pleat in the upper part of trousers for inserting the belt' - although the vocalism in this form is not quite clear.
Comment: In Tab. š̌aq̇I < šaq̇Iʷ with metathesis of labialisation; cf. also Tab. Khiv. šaq̇Iʷ, Ag. Bursh. q̇Iuš, Burk. ʡašʷ etc. There is not enough information to establish the PL obl. stem.
Comment: Cf. also Lezg. Akht. q̇ʷeṭ 'cave'. 4th class in Arch. Obl. base *q̇Iʷoṭɨ- (the -o-grade is probably reflected in the Archi direct base q̇Iuṭ), cf. Lezg. q̇ʷeṭi-.
Comment: Obl. stem *q̇Iʷele- (cf. Tab. q̇Iula-, Ag. q̇Iulu-/q̇Iula-, Rut. q̇Iulí-, Tsakh. q̇Iowa-). 3d class in all class-distinguishing languages. On the Arch. (noq̇Ion) and Ud. (meIl) forms see under *nVq̇Iʷe-l.
Comment: Cf. also Tab. Kand. ʁiǯiq̇, Düb. q̇ijiq̇i, Ag. Fit. q̇Iadiq̇I, Burk. q̇areq̇, Lezg. Khl. q:ciq̇ 'sheep-skin'. In some languages we observe a dissimilative change *q̇- > q:- (Tab., Bud., Lezg.). Interesting is the exceptional preservation of nasalisation in the Budukh form (*q:iǯeq̇ would be expected). Belongs to the 4th class in all class-distinguishing languages. See Талибов 1980, 307.
Comment: Cf. also Ag. Bursh. q:Iaraq:Ial, Fit. q:Iaraq:il, Burk. q:araq:il, Tp. q:äräq:äl. 3d class in Rut. and Tsakh. Phonetically aberrant are: Lezg. kerekul (Khl. kerekel), Kryz. ḳeräḳel 'magpie'.
Comment: In Bud. cf. perhaps also the adjective q̇uq̇al 'spheric, round' (which would make it possible to compare also Arch. q̇ʷerq̇ǝ-bos 'to roll' < *q̇ʷelq̇V-). 3d class in Rut., Tsakh. and Bud. In Udi there is an unclear suffixed -pun.
