COMMENT: Reconstructed for the PEC level. One of the roots with the rare voiced fricative *ɣ. The Lak. form may be ultimately borrowed from Av. (with a semantic change), as are many Andian and Tsezian forms listed above; however, the PN : Av. : PA correspondence can not be explained by borrowing.
COMMENT: For PEC we must reconstruct an intermediate form (with assimilation) *ɣwārdV̆. Some Caucasian source is probable for Osset. ḳord / ḳʷard 'group, flock' (despite Abayev 1958 who suggested to derive the Osset. form from Got. *kaurdr 'herd' - an impossible form - and proposed further the direction of borrowing Osset. > Adygh. A reverse direction is more probable in this case, although phonetic problems still remain). A special problem is the etymology of PL *ḳareṭ 'group, herd' (Tab. ḳereṭ, Ag. ḳaraṭ, Lezg. ḳereṭ): it can not be derived from *ɣwārdV̆ ( > Arch. ʁat:ǝra) and may be, in fact, a secondary loan from Ossetian (Alanian).
COMMENT: Reconstructed for the PEC level. The PL form is reduplicated; its pharyngealization is probably expressive. Since the Av., Lak. and Darg. forms coincide precisely, borrowing is not excluded.
COMMENT: Reconstructed for the PEC level. The root contains a rare combination of *ʁ- and the emphatic laryngeal -ʕ-, which influenced the development of *ʁ; nevertheless, the etymology seems acceptable (despite the irregularity of vowel -u- in Lak.).
COMMENT: Reconstructed for the PEC level. Some irregularities may be explained by lexical interference: -t- in Lak. is probably due to a merger with PEC *tălV ( /*lătV) 'kennel, gutter' q.v. (cf. especially the Khosr. variant aItala); *q: in PD may be due to the influence of PD *q:at:a 'ravine' ( < *G_ĕdV q.v.).
COMMENT: Reconstructed for the PEC level. The PA and PL/Khin. forms are obviously related, although reasons for variation between č̣ and č are not clear.
COMMENT: Reconstructed for the PEC level. An expressive reduplicated root, but correspondences are regular and the PEC reconstruction seems probable (although there is a possibility of borrowing either in Lak. from Darg., or vice versa).
COMMENT: A reduplicated and obviously onomatopoeic root; it is, however, very widely spread and in some cases serves as a basis for the words meaning 'to speak, talk', so perhaps its PEC reconstruction is not completely senseless.
COMMENT: The identity of the Av., PL and Khin. stems is obvious, although it is rather hard to reconstruct the precise function of this pronoun. In PL it is used to denote objects located higher than the speaker; in Avar - rather vice versa (it is either a neutral third person pronoun or denotes objects located lower than the speaker). Cf. also Khin. q:ʷa (with emphatic tenseness, the same as in t:ʷa < *dV) 'there (below the speaker)'.
COMMENT: An expressive common NC root. The etymology seems rather probable, inspite of the irregular assimilation (ʁ- > q̇:-) in Andian.
Abdokov (1983, 117) compares the Ubykh word with PL *q:Iʷen: 'partridge' (reflected in Lezg. q:ʷed, Khl. q:Iʷäd, Tab. ʁIud, Ag. Rud, Fit. ʁIud, Rut. ʁIud, Tsakh. q:Ion, Mik., Gelm. ʁIon, Arch. qIon). The Lezg. root, has, indeed, no other etymology, but the phonetic side of the comparison is somewhat dubious (PL *n: does not usually go back to clusters with dental consonants).
COMMENT: Reconstructed for the PEC level. The Lak. form is an irregular transformation of *halli-bVr = PA *ʁʷaldi-bVr (with an original plural suffix).
COMMENT: It is interesting to note that the roots meaning 'summer' and 'winter' have reversed their meanings in PN. The whole problem of reconstructing seasonal subdivisions in PNC still awaits investigation (there probably existed several major generic terms and a variety of names for relatively short periods of time, just as in modern languages). Despite semantic difficulties, it seems hardly possible to separate the Nakh forms from the rest.
COMMENT: The comparison seems plausible both phonetically and semantically (although the root is not very widely represented). See Abdokov 1983, 106 (Tsez. : Ub.; other EC forms cited by Abdokov, do not belong here).
