I guess I'm sticking with the obvious. We're playing a game, and things should work the way/s the rules say they do, the vast majority of the time. That seems like a standard that should apply, whether a given thing is magic or not.
A year ago now I posted
[1] So far as pre-existing norms extend, participants can often agree that a description D will have the consequences C.
[2] Rules supersede pre-existing norms and extend beyond them.
[3] During play it can be decided if any D has the consequences C by matching that D to a norm or rule that explicitly states or implies that C.
What you are appealing to is the notion that where it says that
backgrounds in this chapter provide both concrete benefits (features, proficiencies, and languages) and roleplaying suggestions
The background "features" concrete benefits are game mechanics that supersede and extend beyond other norms. However, it's also evident that said superseding and extending can fail to banish pre-existing norms. PF and 4e debates over prone snakes provides some evidence of that. And they
must fail because our capacity to grasp rules depends upon a web of meaning that is never completely specified by the rule.
The following may prompt reflections on that (emphasis generally mine)
in a world where magic works the way it's presumed to in most D&D 5e settings, things that seem as though they'd work in-game the same way they would in the real world might not in fact work the same way they do in the real world. In other world, internal setting logic might mean the laws of reality in the game world aren't the same as in the real world, so using the real world as your basis will be an error.
The world needs to behave in ways the players can predict, so they can make decisions on some reasonable basis. Sometimes--mostly in matters of classical physics--this means things in the game world behave pretty much the way they do in the real world. Other times, this means they behave in ways more consistent with the game rules.
These two quotes can be reconciled by supposing the group have meta-norms that guide them between "using the real world as your basis will be an error" and those "sometimes--mostly in matters of classical physics" cases that follow some other norms for fictional-truth determination. Something like David Lewis' principle of least difference from collective beliefs.
I think the players are still bound by the fictional position. I'm pretty sure the player in my third campaign who has that feature isn't expecting it to work outside the starting city--I'm (I hope, obviously) going to make it clear it still will. Hell, she might not think she still has access to it in the starting city (she was granted an Out by the Guilds) and I should make it clear her character can, in fact, still make non-hostile contact and can plausibly at least get information at minimal cost.
As it appears this player is trusted to have self-regulated toward.
While we can talk about "mundane" versus "magical," IMHO, that distinction doesn't really matter too much when we view this from the lens of genre: i.e., fantasy adventure. For example, there are a lot of mundane happenings in the realm of superhero comics, but even the mundane lives of the characters, whether they are superheroes or the supporting cast of normies, are extraordinarily dramatic with a lot of dubious realism for the sake of dramatic superhero storytelling. The same is likewise true for non-magical genres like crime dramas, science fiction, and the like or even Hallmark movies. The genre of fantasy adventure, particularly D&D-style fantasy adventure, is more concerned with its tropes than its realism... I don't really see anything inconsistent between D&D 2014's background features and D&D's genre fiction.
The setting is more important to me than the ruleset, so I make sure the rules can model the setting as closely as I can practically manage.
And it appears that such beliefs include as regards the nature of the fiction itself, i.e. that it's genre or milieu is such-amd-such and this is to prevail over other norms.
Often it's enough to find a phrasing that makes the proposition fitting, such as
the original poster... knows the strict rules answer... Instead, he is looking for advice on how it "should" be played.
Which may lead to
“This condition can affect limbless creatures, such as fish and snakes, as well as amorphous creatures, such as oozes. When such a creature falls prone, imagine it is writhing or unsteady, rather than literally lying down.”
To me the rules by which fictional-truth is determined are subtle and remain puzzling. It seems that there can be a principle in play (which may be unwritten) with sufficient force to bend game mechanics to some other perceived demands. And I feel confident in observing that not all elements of a given fiction must obey the same set of fictional-truth determinants, based on the ability of folk to articulate different sets and speak about choosing between them.