COMMENT: A rather complicated case (as is usual with cultural words). First, we must for phonetic and semantic reasons dismiss the possibility of borrowing from Arab. ḳabāʔ 'langärmeliges Obergewand' (which, however, is also present in Tsezian languages: Bezht., Gunz. ḳaba, Gin. ḳoboj 'shirt'). Second, we must exclude interlingual borrowings: Arch. part:uq̇ 'fur-coat' from Lak. (barq̇ut, dial. bartuq̇ with metathesis); Lak. q:awaču ( < *q:abaču) from Avar or from Lezghian languages; some of the Andian forms also may be borrowed from Avar. Lezghian (Tab., Rut.) forms could be also considered as Avar loanwords. However, there are reasons to reconstruct the derivate *q_ārṗV-čV (or *-ćV) already for PEC, cf. Hurr. qurbiži 'garment worn under mail-coat or helm' (see Diakonoff-Starostin 1986, 18), which was borrowed in Akk. kurbisu,gurpisu. Thus the PL form *q̇ap:a(r)čaj may well be genuine (with a shift of laryngeal features usual for roots containing two or more stops).
COMMENT: Reconstructed for the PEC level. Except for occasional assimilations (usual in a root with two stops), correspondences are regular. Cf. also HU: Hurr. karkar-ni, Ur. qarqara- 'armour' (see Diakonoff-Starostin 1986, 58). It is interesting to note the recurring meaning 'coat of mail, armour' (in Nakh, HU, Lezg.): it shows that the root may have designated some piece of armour or warrior's garment in PEC.
COMMENT: Reconstructed for the PEC level. The stem is violating the rule of "uniform strength": the first consonant is reflected as tense, but the second - as weak. This is obviously due to the influence of two other roots, PEC *GĕlćwV and PEC *bHǟćwa (q.v.), with a very similar meaning.
COMMENT: Reconstructed for the PEC level. An expressive root with several irregularities: the most severe one is PA *χ:- instead of expected *q:-. Other irregularities are mostly assimilations (in PN: *q̇- < *q- before *ṭ; in PD: *-t < *ṭ after *q-) or dissimilations (in PTs, where -ṭ- would be expected). All this is rather usual in a root with two stops (especially in an expressive one).
COMMENT: Not quite clear is velar *k:ʷ in PWC (although there are several cases of such a correspondence; perhaps the explanation lies in the special development of the initial cluster); otherwise correspondences are regular, and the reconstruction seems reliable.
A variant of the same root is, perhaps, reflected in Cham. ʁoc: "rye" ( presupposing PA *ʁʷič:V); reflexes of both consonants are, however, irregular, and we should perhaps think of it as an old loanword (from some related language). The Cham. (*Tind.-Cham.) form was, in its turn, borrowed in Inkh. ʁuč 'rye'.
COMMENT: An interesting common NC root. Despite assimilative / dissimilative processes (typical for roots with two stops), the etymology seems quite satisfactory.
COMMENT: Reconstructed for the PEC level. The vocalism is not quite clear. In PL we have to assume an early process *q:Vṭ > *q̇:Vt. In [Diakonoff-Starostin 1986, 28] we considered Lak. -n- as going back to *-l-, and compared Urart. quldǝ 'steep' (or 'precipice'), but Lezghian data seem to contradict the reconstruction of any medial resonant. Therefore, medial -n- in Lak. must be secondary (on analogy with kuInṭa 'pit, small pit'), and the Urartian word is probably unrelated.
COMMENT: Although the word is cultural, and there was active interlingual borrowing (among the forms listed above Av. q:abá may represent a loan from Nakh, and Darg. qaba - from Avar), there is little doubt in its PNC antiquity; minor phonetic aberrations are easily explained by the root structure with two stops, leading to assimilative/dissimilative processes. It must have denoted a big wooden vessel (hence the meaning 'boat' in PWC), with a secondary development > 'earthenware vessel' (PN, PA). The meaning 'sack, bag' in PL is obviously secondary, possibly under Turkic influence - nevertheless, PL *χ(:)ʷap: can not be regarded as a loan from Turkic *kap (the initial fricative χ- can not render Turk. k-).
COMMENT: A Lak.-Agul isogloss. Lak. pharyngealization is possibly due to the influence of qiIṭ 'beak; adam's apple'. Borrowing is not excluded, thus the PEC antiquity of the root is dubious.
COMMENT: Reconstructed for the PEC level. The root is rather peculiar (although there are no doubts at all in its EC antiquity): it contains a very rare phoneme *-ɵ- (which yields t-like reflexes in PTs, Lak.and PL, but s-like reflexes in And. and Darg.). Several subgroups also reflect a specific variation of reflexes (*qw- or *χw-) which is a rather rare case.
Since the root is liable to metatheses (cf. the situation in Lezg.), one would be tempted also to compare PAK *tħa- in *tħa-(m)bǝLǝ 'lung' (Ad. tħābǝL, Kab. tħambǝL), Kab. tħa-m-ṣ́ǝʁʷ 'liver' with a possible reconstruction PWC *tǝχ́Iʷa or *tǝq́Iʷa; however, it may be a pure coincidence, and Kuipers' analysis (*tħa as 'life, spirit') may be right (see Shagirov 2,83).
COMMENT: Reconstructed for the PEC level. The root reveals some variation of laryngeal features, typical for roots containing two stops, but in general correspondences are satisfactory. Some other languages reveal similar but irregular forms: cf. Av. ʁud, Arch. χIit 'spoon' - perhaps, old interlingual loanwords.
In PL we would rather expect *q:uṭ, not *q:ʷɨṭ (reconstructed on basis of Kryz. q:ʷaṭ-ɨǯ). It is, in fact, possible that the Kryz. form reflexes a merger with another root, reflected mainly in the Av.-And. area: cf. Av. ʕortó 'a cylinder-shaped vessel' (whence probably Lak. uIrtu 'clay vessel'), PA *q̇ʷVrtV / *ʁʷVrtV > Akhv. q̇oto (Tlan. q̇ortu, Ratl. q̇orto, Tseg. orto) 'a measuring vessel', Tind. ʁʷarta 'mortar'. In that case a reconstruction *q:uṭ is also possible for PL.
