> > The form vecos is found in Old Latin (CIL388, plural at CIL1806).
> > The process is:
> > *woikos > *weikos > we:cos > vi:cus
>
> Are you implying that Old Latin <e> represented 3 sounds: /e/ >
> Latin short /e/, /e:/ (presumably open) > Latin /e:/ and /e:/
> (presumably close) > Latin /i:/? Old Latin ought to have had /e:/
> that did not change to /i:/, e.g. *ve:ros 'true' (Pokorny root
> #2172).
The spelling VECOS is what is attested. What it stood for can be debated.
Ernout (Recueil de Textes Latins Archaiques) simply says "with reduction of
the diphthong". So I guess one option is a pronunciation /ei/, with
spelling <E>. Maybe this is more likely.
However there are a number of proposed eytmologies in which inherited /e:/
appears to go to /i/, for example:
fi:lius, suspi:cio, convi:cium, subti:lis, vi:lis
and deli:nio, doublet of the commoner dele:nio.
Not all are widely accepted, but vi:lis at least would be nice to
believe.e.g.:
vi:lis < *we:lis < *wes-li *wes as in ve:num < *wes-no, RgVeda vasna
price, Hitt wa-a-s'i "buys"
So i: < e: < *ei/*oi is not impossible; but I take your point, that if
*ei/*oi pass to *e: which is identical with the inherited long /e:/, the
outcomes should not be so regularly /i:/ against /e:/ from *e:
Peter