Just because you are not obliged to roll the stats of every single NPC if their stats aren't going to get used in the story, this in no way means that such NPCs don't HAVE Str/Dex/Con/Int/Wis/Cha scores!
It means they don't exist, in a sense. Tree-falling-in-the-forest stuff, only, hey, if I say, there's a city off on this edge of the map, and the PCs never go there, then, in a sense, the hypothetical inhabitants 'have not stats.' They also have no names, and, in essence, don't exist. Afterall, they don't exist, anyway, even if the party did visit the city, and I did determine stats for a few of 'em.
Quote the 5E text which says that the general population is not assumed to be consistent with the 3d6 bell curve.
It's not a fair task actually, because no book is going to do that. But what they would do, if they wanted to change that assumption, is tell us what the new assumption is.
Exactly, there's no population demographics given with respect to stats. The closest thing is the vague implication of straight-10 commoners and more capable NPCs of other sorts, and it's not inconsistent with the bell curve.
Only if you believe that the rules from previous edition still apply to the current edition unless specifically overwritten. I don't, it's not a written assumption anywhere.
If they liked it so much, they included it in the current rules.
Thing is, not everyone liked every bit of every past edition, so leaving vast swaths of assumptions unstated left room for even the weirder preconceptions.
Hps, for instance. If they'd come down hard on a hp interpretation, it'd've been Edition-War II.