PAN *S was lenited and lost (via *h) in Malayo-Polynesian. Earlier
reconstructions of PAN, before the role of the "Formosan" branches
was sufficiently appreciated and taken into account, suffered from a
strong MP bias (comparable to the Sanskritic bias in early PIE
reconstructions). Of course you can choose the most convenient one
out of a host of forms at your disposal, but there is a high cost
attached. If your choice is arbitrary, that invalidates your argument
(without formal rigour, no logical demonstration is possible). If you
commit yourself to a concrete choice (e.g. Proto-Malayo-Polynesian),
that in turn restricts the time-scale of your scenarios. PMP emerged
ca. 3500-3300 BC, definitely too late to provide PIE with its low
numerals. Then, even *(h)epat- (let alone much younger forms quoted
so freely by Manansala) would be a poor prototype for *kWetw(o)r-.
Piotr
--- In cybalist@..., tgpedersen@... wrote:
> Life is hard, you can't be happy very long (sigh).
> As I recall, you had a derivation PAN *Sapuy > *hapuy in an earlier
> posting?
> That's the great thing about proposing loans, you get to choose the
> most convenient form available (within reasonable limits, of
course).