Notes :Since the word is found with *č- in North Khoisan, there is no need to suggest borrowing (unless it is from a more rare direction, North Khoisan > Non-Khoekhoe).
Notes :NK forms are sometimes thought to represent an old borrowing from Khoe, but the possibility of common inheritance is equally strong (the word is firmly reconstructible for North Khoisan, and, on the other hand, not found in Nama - the most obvious source of borrowing - in the meaning 'sun'). If the Sandawe form belongs here, what we have is quite possibly a bisyllabic structure reduction with secondary click formation: *cVmV- > *cVm- > *|Vm-.
Notes :Possible development in Sandawe: kiso- < *kso- < *cko-? The parallel is somewhat forced, but the ultimate solution depends on perfection of the intermediate reconstructions. On a weird sidenote, cf. S. Starostin's reconstruction for Proto-Sino-Caucasian: *(t)xq̇wV̄ 'two'. Old "global root"?
Notes :The antiquity of the South Khoisan root depends on whether |Xam dum 'shadow' belongs together with the rest of the forms. The Sandawe parallel is weak (we would not expect a glottalized click to develop into plain *s-).