On 2008-11-11 20:42, tgpedersen wrote:
> You're right I'd have to appeal to analogy.
> So would you with your composite class VI.
Yes. The difference is that I don't attempt to explain the original
*a/*o:/*o:/*a set with recourse to unknown words in an unknown language.
I say "unknown" because the historically known Venetic language did not
have any of the properties you ascribe to your ubiquitous substrate. In
particular, it was not an ablaut-free a-language -- that much is clear
enough.
Piotr