Comments: PAT *ṭʷǝ-ʕʷa; PAK *bźa-q:ʷá (the etymology of *bźa- see under *bǝ̃źʷa). Shagirov (1, 88) thinks that -q:ʷa here is the same as in *ʎa-q:ʷá 'foot, leg'. This could be true if the compound had meant originally 'toe-nail' (lit. "foot-horn"), cf. an analoguous semantic development in Balto-Slavic (*noga). The same root may be discovered in other derivates like Kab. q:ʷā-gʷa 'hornless, one-horned', Ad., Kab. q:ʷā-pa 'corner, angle', Kab. q:ʷā-ps 'root' etc.
Ub. q́a and PAK *q:ʷa point unambiguously to PWC *q:́ʷa; however, in PAT we would rather expect *qʷa than *ʕʷa. The reason of the irregularity is that the PAT form must have originally been a compound: 'two horns', PWC *tqI:́ʷǝ-q:́ʷa which underwent a dissimilation and was transformed into *ṭʷǝqI:́ʷa > PAT *ṭʷǝʕʷa. In this way the numeral retained here the initial dental, which was lost in the independent position (PAT *ʕʷǝ 'two' q.v.).