> Maybe Glen has already commented on this, but it may
> be noteworthy that some are tempted to see *ok^toH (or
> whatever) as literally 'twice four' and *nowm. 'nine'
> as being related to the very similar root *new 'new'.
I did actually. I suggested my own alternate theory
that *okto:u is indeed "fours" (plural rather than
dual though) and was Mid IE *kWetWa-xe (early Late IE
*ktwax > *aktwax, due to the a-prothesis rule that
also affects the development of *osdos "branch, perch" <
*sed- "sit" and *oxwiom "egg" < *xawi- "bird").
However, I still don't think that numbers under ten
were unknown to them. In fact, I think that "ten" is
reconstructable far back.
> This is, however, an argument that they ought to have
> been borrowed together.
But still, even with a plea to numerology, it doesn't
necessarily mean that the numbers were borrowed at the
same time, since these mythological concepts endured for
centuries, if not for millenia.
However, as I told Miguel *septm would have to be from
Common Semitic since East Semitic doesn't work to explain
the mimation chronologically or geographically. However,
I can accept a NWSemitic *s^idc^u having been borrowed a
little later, maybe a matter of a couple of centuries.