Comments: There are two PWC roots, which are probably later differentiated variants of one original root for 'ear':
a) *ŁA- in compounds meaning 'ear': PAT *lǝ-mǝħa (where -ħa must be the same root as PAT *ħa 'to hear' q.v.; see Ломтатидзе 1955, 824); PAK *tħá-ḳʷǝma and Ub. Lá-ḳʷǝma (etymology of the second component obscure);
b) *ĺA- in compounds meaning 'to listen': PAT *-ʒ́ǝ-rǝ-ʕʷǝ- (Abkh. á-ʒrǝʕʷ-ra, Bz. á-ʒ́ʕʷǝr-ra /with metathesis/, Abaz. ʒǝrʕʷ-ra); PAK *da-ʡʷa- (Ad., Kab. daʡʷa-n); Ub. La-q̇ʷǝ́-. Here Shagirov (1, 150) thinks that PAT *ʒ́ corresponds to PAK *d - this is certainly wrong, because in this case (PWC *b́ʷ) we should expect dʷ, and not L in Ubykh. PAT *ʒ́ǝ- in this verb is probably a modification of the reciprocal version prefix *-źǝ- (which is semantically rather plausible). The element corresponding to PAK *d and Ub. L here must be PAT *r (*-rǝʕʷǝ-) which (despite Shagirov ibid.), of course, can not be considered a causative prefix (the verb has no causative semantics at all). For the second part of this PWC compound see *q:́Iʷǝ-. The root *da- ( < *ĺa-) is present in another PAK compound *da-gʷǝ́ 'deaf' (where -gʷǝ is a privative suffix: "earless"), see Шагиров 1, 145. Judging from the external evidence, the variant with the resonant *ĺA is original; reasons for its transformation into a lateral affricate in the word for "ear" are yet to be investigated.