> 1) The Greek attestations of the name, which mostly end in -E, or -AI
> which was pronounced like a long "e".
>
> 2) The hypothesis that the name is derived from a form of the verb HWH,
> which would typically end in -e(h).
ai-e could be just a suffix of duality, modeled upon plural Elokim.
Clause 2) cannot explain e-ai shift, while the reverse shift, dual ai-e, is
likely.