Notes: An expressive (and rather universal) root; but there is no doubt in its PNC antiquity. Reduplicated forms like Lak. k:urk:ur or PAA *gʷǝrǝgʷǝrǝ may provide a link between this stem and the reduplicated *gīrgwV q.v. See Абдоков 1983, 146.
Notes: An expressive root (which is sometimes hard to distinguish from a similar *qwɨ̄ṭi q.v.; partial contamination has occurred). Pharyngealization in PL is clearly secondary (expressive) - otherwise PL *qIʷ (not *q:Iʷ) would be expected.
Notes: Reconstructed for the PEC level. Correspondences are regular.
Cf. also Osset. qīs / ɣesä 'thick hair' < Iran. *gaisa- (see Abayev 1973, 304-305). If not for the labialisation (reflected in PA), it would be tempting to consider the listed Caucasian forms as an old loan from Iranian.
Notes: A rare root, preserved only in some periferal languages (for the discussion of PA *ʁʷVčV-n see *qărć̣wV̆). However, the correspondences are regular (except for a metathesis in Khin. - not unfrequent in roots with two stops), and the comparison seems safe. Note an identical suffixation in PA and Khin.
Notes: Phonetically the EC-WC parallel is quite satisfactory. As far as the meaning is concerned, for PWC we have to suppose a change *'dense' > 'thick' which seems to be quite motivated.
Notes: Except secondary assimilations (typical for roots with two stops; PN *gub < *guṗ; PL *ḳaṗ-al < *k:aṗ-al) correspondences are quite regular. The comparison is quite satisfactory both phonetically and semantically.
Notes: Cf. also Ag. ʁeb/ʁäb (Shaumyan) 'haystack', borrowed from Darg. There is a good correspondence between Av.-And., Darg. and Ub., although the vocalic reconstruction is not certain.
Notes: An onomatopoeic root (like many other birdnames), which explains the presence of phonetic variants (voiceless *qw is reflected in PWC, voiced *Gw - in PA and PL; PN *q may reflect both variants). Cf. Abdokov 1983, 121. Similar onomatopoeic roots are attested in other languages (cf., e.g., PK *ʁuʁun- 'to coo').
Notes: Reconstructed for the PEC level; the common meaning can be formulated as "some clothes for the upper part of body". A cultural root; similar forms can be found in many languages, and PEC may have borrowed it from some neighbours. Cf. Indoeuropean forms: in Slavic - Russ. куpт(к)a, Ukr. куpт(к)a, Pol. kurt(k)a (despite Vasmer 2, 429-430 etc., it is hardly borrowed from Lat. curtus 'short'); in Indo-Iranian - Osset. kʷǝrät / kurät 'shirt, jacket', Sogd. qwrty 'shirt', Avest. kǝrǝti- 'a k. of clothes', Old Ind. (late) kuratu- 'shirt' etc. (see Abayev 1, 609); in Kartvelian - Georg. ḳwarti 'shirt', Megr., Chan. ḳurta 'drawers, pants'; in Turkic - Turk. kürtä 'short dress' etc. (Radlov 2 1265, 1461). A direct Iranian source should be suggested for Arm. grat 'war cloak', whence Ud. gurat: 'shirt'.
Notes: Besides the listed languages in EC the word is present also in Lak. (kutan), Av. (kután). The Lak. and Av. word are probably borrowed (cf. the secondary devoicing of k-, resembling the devoicing in the PL form: it is, however, hardly possible to deduce PL *kʷitan from Lak. or Av., thus the loan was probably reverse).
The PNC reconstruction given here is highly tentative, primarily, because of the irregular correspondence between EC and WC. The word is also present in other languages of the Caucasus: Georg. gutani, Osset. gūton / goton, Arm. gutan, Azerb. kotan, but does not have a secure etymology in any of these languages. See Genko 1930, Abayev 1, 527: "a common Caucasian word of unknown origin".
Notes: The Darg. form is somewhat dubious (not clear is the initial uvular instead of an expected velar); otherwise the correspondences are regular, and the root may be safely reconstructed for PNC.
Notes: A Lak-Lezg. isogloss. There exists, however, a HU parallel: Hurr. qund-arǝ 'mountain, abode of gods' (see Diakonoff-Starostin 1986, 24), which makes the EC antiquity of the root rather probable.
Notes: The comparison is reliable both phonetically and semantically. There is some uncertainty about the vocalic reconstruction (since there are general doubts about the vowel reconstructed as *ü).
Notes: The root is very sparsely represented; the Rutul word may belong here if the original meaning was 'lean; bad' (with a typologically frequent change > 'left'). Not very secure.
