--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
> ... For Latin, you'd have to assume a completely isolated survival
> of a purely hypothetical periphrastic stage, since *<porta:mini:
> estis> is not attested; nor is *<porta:minus>.
Latin does show a few fossil survivals of a mediopassive participle in
-menos, e.g. alumnus < alomenos, and femina = "woman" < dhe:-mena: =
"she who is suckled from".
I suspect that one of the various threads leading to the Latin perfect
in -vi is a collapsed and analogically levelled form of a periphrastic
form with the -wo:s- participle, e.g. amavit < *{ama:wo:s esti}. That
would explain why Latin perfect passives such as **amavitur never seem
to be found, as **{ama:wo:s estur) or similar would have no meaning.
As an example of languages losing participles, Modern Greek verbs have
kept only 2 of the many participles that Ancient Greek had (except
likely as artificial revivals in the `katharevusa' dialect).