Notes: The root is expressive and subject to irregular changes (metatheses etc.), therefore the reconstruction is rather tentative. Some other EC forms can be regarded also as very irregular reflexes of the same root: cf. PD *qisqa (Ak. qisqa, Chir. qisqa-ne etc.), Bezht. (Khaid.) kala-ʁasq̇i (with further transformations: Gunz. kɔlɔq̇i / halaki) 'spider'; see also comments to Andian and Lezghian reflexes.
Despite all these difficulties (rather expected in a word meaning 'spider'), the root seems more or less reliably reconstructable for PNC.
Notes: Reconstructed for the EC level. The root is close to PEC *qɨ̆lʔi 'arm; wing' q.v. and both forms have partially merged together in some languages.
Notes: Reconstructed for the PEC level. A similar form exists also in Georgian (χaχwi 'onion') and Ossetian (xaxwɨ, see Abayev 1989, 149), where it is most probably an EC loanword.
Notes: At least the Lak-Darg. parallel seems reliable (the And. form is known only from Khaidakov's dictionary, which is a rather unreliable source), but the original meaning is not quite clear: the meaning in Lak. ('raw hide; membrane') may be primary, or else it may be derived from 'bat' ('bat-like skin'). The word for 'bat' is not recorded in most EC languages, and we must wait for further data.
Notes: An expressive reduplicated root. Correspondences are, however, regular (except for a usual assimilation in PA: the expected *χ:orχV > *χorχV), and the PEC reconstruction seems probable.
Notes: A Nakh-Darg. isogloss with somewhat irregular correspondences (due to the root's expressive nature). The specific cluster *št (*šṭ) in PN is probably a result of suffixation (*χHVrć̣-ṭV; cf. an analogous dental suffix in other 'weasel' names: PN *dingV-ṭ,*šVrṭq̇V < *šVrq̇-ṭV).
Notes: Reconstructed for the PEC level. Inspite of some minor problems (due to the expressive nature of the root), the reconstruction seems rather secure. Cf. also Hurr. χaχli 'cheek' (see Diakonoff-Starostin 1986, 61).
Notes: Reconstructed for the PEC level. The correspondences between PN, Av. and Rut. are fairly good. The PTs form poses some problems: we should rather expect *χ:ɨχ:V. We must probably reconstruct *χɨwχV for early PTs, with a development *χɨwχV > *χ:ɨwχV (with regular lack of intensification after a resonant) and subsequent assimilation *χ:ɨwχV > *χɨ(w)χV.
Notes: A common NC verbal root with a rare structure CVRCV. Despite numerous irregularities (variation between *ć and *ś, secondary voicing in PWC, metathesis in Khin.), it is difficult to separate the listed forms.
Notes: Reconstructed for the PEC level. Correspondences are regular. EC > Osset. xal / xalä 'grass stalk' (see Abayev 1989, 134; see ibid. some Georg. and Slavic parallels).