Not that I would disagree, but what happens if we assume with Cuny
and Glen Gordon that what is traditionally reconstructed as PIE <o>
and <e> was actually <a> and <&>? We'd then have for Indo-Iranian
<&> -> <a> and for Greek etc <a> -> <o> and <&> -> <e>. Suddenly II
looks closer to PIE than Greek etc.
Torsten