I'm willing to consider that.
Since there is a root *h₂melǵ- (to milk, to press out / suck etc) which gave Attic amélɡɔ, Latin mulgeō, Old English meolc (English milk) etc
Maybe it could have been related to that.
If we assume that the reconstructed root is correct, we need a language that shifted h₂ >a (like Greek), e > a (like Sanskrit) and ǵ > t
h
and a suffix like ancient Greek -εια
https://en.wiktionary.org/wiki/-εια#Ancient_Greek
Now, concerning the last there are words where before front vowels (e, i) there is a dental in Attic /th/ (MGr /θ/) that derives from a reconstructed *gʷʰ (l'm leaving my personal opinions about reconstructions aside)
Example:
https://en.wiktionary.org/wiki/Reconstruction:Proto-Indo-European/gʷʰer-
Here we don't have *gʷʰ in the reconstructed root, though.
So, for now, I can't say that it descends from that root.
A thought experiment:
h₂melǵ-eia would give ΑΜΕΛΓΕΙΑ in a dialect like Attic.
In a language, which was similar to Attic but in which e shifted to a we would have (e shifted to a in Sanskrit, for example)
AMAΛΓΕΙΑ
In a language, like the one above, that also shifted ǵ to 'dh' (which happened in Albanian) we would have
ΑΜΑΛΔΕΙΑ (but Δ would represent ð)
{ΑΜΑΛΘΕΙΑ is supposed to have been pronounced amalt
heia according to the reconstructed Attic pronunciation.
t
h shifted to /θ/ in Later Greek (also to an
sound of some short in some Doric dialects early on)}