[tied] Re: sum

> > All Jens had to say was that the semi-thematic paradigm is rare.
> > Schmalstieg quotes Meillet as saying that comparative grammar

should

> > use anomalies, ie. survivals, rather than regular forms. I

second

> > that.
>
> That can hardly be extended to an anomaly unsupported by data from

other

> IE languages, and one for which a simple and plausible language-

specific

> explanation is available. Otherwise you would have to insist that
> irregularities like go/went and person/people must be survivals.

I don't insist and I don't say 'must'. I always say 'might'. You're
confusing your own habits with those of other people.

> The
> curious alternation of *e/*o in the thematic paradigm is a

peculiarity

> shared by several branches and so unlikely to have developed
> independently. The semithematic pattern, on the other hand, is a

local

> oddity best explained as a local innovation.

'Best', why, except by your preference?

>What generalisation would
> have produced *bHer-e-t/*bHer-o-nt out of **bHer-t/**bHr-o-nt?

*bhr-o-nt > *bher-o-nt by regularisation of the root. *bher-t >
*bher-e-t by regularisation of the suffix.

>To ask
> the most obvious questions, why is the "generalised" thematic

vowel *e

> in the allegedly innovated forms if the only models for its

insertion

> had **-o-?

By Jens' rule, whatever its phonological cause, that -e- is
preferred before unvoiced sound.

Why does the *e/*o alternation (certainly a non-trivial

> correspondence) recur in branch after branch, if it's post-PIE?

As it would be if the semi-thematic paradigm (as per your fiat) was
local to Latin.


Torsten