On 2007-02-07 15:16, Jens Elmegård Rasmussen wrote:
> Wouldn't you rather wait with the voicing, making the chain of events
> *-sr- > *-Tr- > *-fr- > *-Br- > -br-? I think this makes it closer to
> the word-initial changes.
Both would work; the changes are parallel and natural-looking in either
case. I like, on grounds of overall economy, the Hartmann/Rix theory of
voiced aspirate treatment in Italic (no phonetic devoicing medially),
hence my preference for a model of proto-Italic phonology in which the
distribution of voiced and voiceless fricative allophones is more or
less as in Old English (no voiceless ones between sonorants). But it's a
matter of personal taste rather than certainty.
Piotr