Комментарии:Suffixes do not match in Kui and Kuwi. Since -nḍi is a much more rare and less productive suffix than -ṛi, I think that the protoform should be reconstructed as *lapenḍi. In Kuwi_Mah a metathese: lapeṛi > ṛapeli.
Куви (Фитцджеральд):lecali (let-) "to hang up, deck oneself with flowers"
Куви (Шульце):leh'nai
Куви (Маханти):rec-
Куви (Израэль):lej- (-it-) "to hang (intr.)"
Дополнительные формы:Also Kui ṛaki hung, hooked; ḍēnga (ḍēngi-) to be suspended; Kuwi_F lepi kīali to deck (another) with flowers; lepi aiyali to be hung; Kuwi_Isr leʔ- (-t-) to hang (tr.)
Комментарии:A complicated case. First of all, ṛ- is not the regular Kui reflex for PK *l- (ḍ- should be expected, and indeed we find it in ḍēnga). However, there is at least one more example of PK *l- > Kui ṛ- (*lapenḍi 'chin' (5181)). It is not excluded that this development regularly happened in Kui before short vowels, since there are almost no counterexamples (most of these have secondarily shortened vowels in Kui). Second, the auslaut of the root is unclear: Kui and some Kuwi forms point to a vocalic auslaut, while Kuwi_Isr intransitive lej- should be traced back to *lej-. However, the transitive-causative form seems to have a stable *-e- auslaut.
Куи:ḍrēnḍu (pl. ḍrēṭka), ḍrēngoni, ḍrēnguni, ḍrēonji, ḍrēoli, ḍrēlu
Кхуттия куи:ḍēḍru (pl. ḍrēṭka)
Куви (Фитцджеральд):lengūńi
Куви (Шульце):lenguni
Сункараметта куви:leŋguṇi
Куви (Израэль):leŋuṇi
Комментарии:The form is confirmed by Pengo-Manda material. However, Kui forms are probably the most terrible of the lot. The regular reflex should be *ḍenguni.
Комментарии:Kuwi_D and Kuwi_P probably reflect an earlier variant of this bisyllabic root, namely, *elimbu. This is quite understandable for Kuwi_D - an extremely archaic dialect which has preserved old bisyllabic roots in a much better state than others, but unexplainable for Kuwi_P; we may suggest dialectal interaction. Inlaut -mp- in some dialects is not quite clear as well.