> > It seems he doesn't agree with you on assigning
> > original root stress as you do (apart from bereté) to this class.
> > What is your source for Old Russian bèru?
>
> You seem to have Stang at hand. The answer is probably there. Check
> under "c. mobile stress".
Here is how Stang reasons:
"
In view of the general impression afforded by the Old Russian texts,
and especially by the Čudov N.T., with regard to the distribution of
end-stressed and root-stressed forms in the 1. p. sg., the question
arises whether the cases of stress on the first syllable in the
recessive type are not due to a late levelling, of the kind we find in
nouns (O. R. ná nebesexU; mod. Russ. ná smertI, etc.).
If this is the case, the question next arises whether root-stress in
the 1. p. sing. might not be of proto-Slavonic origin, and a relic
from a period when the marginal paradigm was mobile. If this is so,
this verbal class would agree with the Slavonic principle that
marginal end-stress in one form presupposes stress on the 1st syllable
and falling intonation in another form of the same word.
...
It seems reasonable to assume that this mobility consisted originally
in an alternation between stressed first syllable in the whole
singular and stressed final syllable in the other forms. This would
best agree with what we find in Sanskrit émi, és.i, éti : iváh.,
itháh., itáh., imáh., ithá).
"
Why is this reasonable? Why always use the neat Skt. primary endings
as a model? Elsewhere Stang stresses the importance of the secondary
endings of eg the aorist for determining the pattern of the Slavic
*present*, which makes sense to me, given their greater irregularity.
Torsten