Is declension too speculative.

I'm not sure if the Greek accent alone is enough evidence. Different ablaut grades would also be valuable. It's striking that both languages only have a long grade, that seems to indicate that something else is going on. Otherwise, they would have to extend the nominative singular grade to the entire paradigm, which is possible I suppose, but quite a change.

So, you'd add a Narten grade and what else? I'm not sure I see any options for the long oblique stems besides leveling, substratum borrowing, or a sneaky *h₁ in the stem. The oblique oxytones from monosyllabic nouns in AG are somewhat uncommon, as evinced by Newt having to add a parameter obloxy= to an early version of {{grc-decl}} to deal with πούς and θήρ. Speaking of *ǵʰwer- and θήρ, which would act as a fair comparison for this situation, what do we think there? I will need to look at my reference material later to see what Beekes and De Vaan say about them.