Does it need to be asserted? A Cosmology with a prime-material multi-verse seems to positively beg for the assumption.
<snip>
the 1e DMG had conversion rules to take D&D characters into GW or vice-versa.
Well, if it's not asserted it's probably not canon!
And the DMG conversion rules for GW are a good example: I don't think [MENTION=37579]Jester David[/MENTION] or [MENTION=7635]Remathilis[/MENTION] argues that, because of those rules, Gamma World is part of the "D&D canon" they refer to. I think they recognise that those rules were addressed to GMs as a device for running their games; that they were not intended to be treated as part of a guidebook to the D&D world.
DDG was the same - the various pantheons were presented as options for authors (GMs) to choose from - not as chapters in a cosmological/theological atlas of the D&D world. It is only MotP that takes it in this latter direction.
And similarly for the alternative prime material planes: they are a world-building device, so you can include new worlds, cross-overs, etc
if you want to. But there is no assumption that, by default, everything anyone ever thought up in D&D world building is part of a common fictional world. This latter, "multiversal" idea - whereby, in playing GH you are
ipso facto also making FR part of your gameworld, because "Greyspace" links to "Realmsspace"; or that, in playing OA you are
ipso facto making Elminster part of your gameworld, because Kara-Tur is a part of the FR - was a subsequent invention. (Springboarding, as I said, of the MotP approach.)
Hence: someone who disregards this subsequent invention, and approaches the catalogue of settings, and even the catalogue within a single setting book, as a list of options and devices for a GM to use in world-building, rather than as a canoical accont of some "Platonic" fantasy setting, is not departing from the way D&D is meant to be played.