>The evidence of comparative linguistics is, if anything, stronger when it
>operates between languages than within a single branch.
That depends what we are trying to prove. If we are trying to prove
existence of a particular root or form, and reflexes of that root or form
are found in different languages, then that is stronger evidence than
finding them just in a single language. But if we are trying to prove a PIE
pattern involving more than one form, we have to find reflexes of more than
one form within the same language. Otherwise we only know that some
languages show this form, and some that - we can't prove that the
alternation was a regular patterning in PIE.
Peter