The opposition of *ʒ́ĭ 'self' (1-2 cl., animated) : *čŭ 'self' (3-4th cl., unanimated) is well preserved in PL. Most other languages have preserved only one of these roots. The development *č- > c- in Lak. is irregular (typical for pronominal morphemes, cf. *čwi 'who, what' > Lak. cu-, ci-). It is unclear also, whether we should relate here PN *šā '(one)self' (Chech. šā, Ing. še, Bacb. ša-jrwa) - phonetically absolutely irregular [perhaps it should be rather compared with PL *šä- 'demonstr. pronoun, that']. The semantics of the PAT subject version is well explained if we assume its origin from a pronoun; the phonetic side is quite satisfactory.