Re: [tied] Re: Bader's article on *-os(y)o

Miguel:

> No, it doesn't _have_ to. Even if *h3 was voiceless, we'd
> have:
>
> *-eh1 > -e:
> *-eh2 > -a:
> *-eh3 > -o:
>
> It's *h3 either way. That's the beauty of it.

There's no evidence for the dual in *h3, period.


= gLeN