COMMENT: Reconstructed for the PEC level. PEC *-l- is proved here by Lak. dialectal forms as well as by the old Lak. loanword in Archi: Arch. qʷal 'udder'.
COMMENT: An Av.-Lezg. isogloss, rather probably having an old Iranian source: cf. Iran. *ava-kana- 'dug-out' > Osset. ogän (wogän) 'a hidden place for supplies' (see Abayev 1973, 225). Avar bugun < *wugwVn is thus very close to the original form.
COMMENT: Reconstructed for the PEC level. The Tsez. word for "goat's wool" (ʁIiṭa) perfectly fits phonetically; semantically it must be explained as *'stalk' > 'long hair' > "goat's wool" (the variation between "hair" and "bush, shrub" is observed also in several other cases).
COMMENT: The comparison seems plausible both semantically ('wolf':'dog') and phonetically; not quite clear, though, are the laryngeal features of the first consonant (we may deal with an assimilation in PEC, and then we will have to reconstruct *kwăǯē; or it may be a dissimilation in PWC: *gwăǯē > *kwăǯē > PWC *kʷǝč́:ǝ).
COMMENT: Reconstructed for the PEC level. Semantically close forms, but known only in four languages and, though phonetically plausible, involving a metathesis - thus not very reliable.
COMMENT: Reconstructed for the PEC level. Phonetically correspondences are quite regular. The meaning 'knee, elbow' in PN is probably secondary ('joint' < *'spine joint' ?).
COMMENT: A difficult case. See Dzhavakhishvili 1937,613-617, Klimov 1963,195-196, Shagirov 1977,159 (comparing the above forms and Kartvelian *katam- 'hen'), Abdokov 1983, 116.
The listed forms are related with a great degree of probability, but several assimilative processes must have happened in different languages (which is usual for root structures of this type, containing two stops. In PN (where *gw- regularly > *k-) *gwāṭa > *kōṭa- > *kōta- (progressive assimilation); in PTs *gwāṭa > *gʷɨṭV > *gʷɨdV (also progressive assimilation). In PL *gwāṭa > *ḳ(ʷ)aṭa (regressive assimilation); on PWC see above.
A suspicious feature is loss of labialisation in PL and PAK; it may suggest that the protoform should be rather reconstructed as *gāṭwa (with shift of labialisation in PN and PAT), because labialisation of dentals is in general easier lost than the labialisation of velars.
In view of everything said it seems probable that the Kartvelian name for "hen" (*katam-) is an old Nakh loanword - since the Nakh form can not be separated from other EC data.
COMMENT: A probable common NC etymon, denoting some domestic equid. (Trubetzkoy's comparison - see Trubetzkoy 1930, 277 - of EC forms with PAK *bzǝ 'female' should be rejected for phonetic reasons).
COMMENT: A Nakh-Tsez isogloss; borrowing is excluded (for phonetic reasons), thus the PEC reconstruction is probable. We must note also the existence in PN of another similar root: Chech., Ing. gowr 'horse'. This could be a loan from PTs *gurV, and thus an etymologic doublet with PN *gile.
COMMENT: Reconstructed for the PEC level. Spread mostly in the Ando-Tsez. area; the Chech. parallel is not very secure (because of lacking Bacb. match).
COMMENT: Reconstructed for the PEC level;correspondences are regular. This is an interesting word, signifying both a beast (hyena, fox, wolf) and a bird (hawk, falcon) of prey.
It is interesting to note that the word for "hawk" in Gunz. is gudba-q (derived from gud-ba "fowl"), and in Abkhaz we have a-ḳʷǝṭ-rǝ́-χ́š́ 'a hawk hunting for fowl' (á-χ́š́ 'hawk') and a-ḳʷǝṭ-rǝ́-bga / a-ḳʷṭǝ́-r-bga 'fox' (á-bga 'fox'). In both cases we have forms containing PNC *gwāṭV 'hen' (q.v.), and phonetically close to the root *gwērdV that we are dealing with. This may have several explanations:
a) a pure coincidence (which is rather unlikely);
b) the root *gwērdV meaning 'beast or bird of prey' merged in Gunz. and Abkh. with *gwāṭV (Gunz. gudo, Abkh. a-ḳʷṭǝ́) and came to mean 'bird of prey hunting for fowl';
c) the PEC root *gwērdV could be assimilated from *gwērṭV < *gwēṭV-rV-, representing the oblique base of *gwāṭV 'hen', and was itself actually meaning '(one hunting) for fowl' in PNC.
[One may note also an aberrant form: Bezht. Tlad. kordaḳo 'kite' which may represent a contamination of this root with another one, reflected in Bezht. Khosh. korzaḳo id.].
We are inclined to accept the third hypothesis (because it explains well all the semantic diversity in modern EC reflexes).
