Latin /a/

[Peter &/or Graham]

Miguel spoke of
>the
>irregular Latin form canis (*k^an-).
And gave a connection to explain the /a/.

I would be interested in other suggestions.   There are a number of words
where Latin has an /a/ which does not appear to be from a syllabic laryngeal
(e.g. where it corresponds to Sanskrit /a/ rather than Skt /i/), and where
it does not appear to be from h2e (e.g. where it does not correspond to /a/
in Greek).

Some occur before /n/, so may be from syllabic /n./ somehow, although the
reason for the syllabicity escapes me at times (e.g. madeo & mando = to chew
[not man-do = to command]; pando, prandeo, scando, langueo).

Others occur near /v/ (e.g. faveo, lavo, caveo, paveo).

Others do not occur near a resonant (e.g. capio, scabo).

Any ideas?

Peter