Congrats for finding an edge case exception. Not that it proves anything.
How is it an edge case? In a mediaeval campaign few people will have grown up in stone buildings, especially if you treat "advantages" like
noble birth and the like as benefits to be rationed by the GM.
And how does it prove anything less than your suggestion, upthread, that a character who grew up in a desert wouldn't know much about trolls? If my example is an edge case, then why is yours not?
My real point is that the notion that there is "natural" or "inevitable" knowledge - like how to find traps in stonework - and then there is "special" knowledge which it wouldn't be reasonable for a starting PC to know - like the vulnerability of trolls to fire - is hopeless. In the actual practice of gameplay, this is all
player knowledge - D&D players
know, for instance, that traps in stonework figure prominently in the game, either because they've read the books or they've been brutally educated in a 1st level dungeon - and it gets imputed to the player's PC.
Imagine a RPGer whose first game was Classic Traveller. Doors, traps and all the other paraphernalia of D&D dungeoneering play no role in Traveller, and a player could be a first-rate Traveller player but be very unskilled in a D&D game because unfamiliar with the tropes and expectations of dungeon exploration. That wouldn't in any sense make his/her PC unrealistic or unreasonably ignorant!
What knowledge checks? If I can have an uncle that knows about monsters that I have knowledge of, I can have an uncle that knows about monsters that I as a player do not have knowledge of and that he has told my PC about.
I have no idea where you're pulling this from. Self-evidently you're not describing your own opinion of the situation. And you're not describing anything that I, or [MENTION=5142]Aldarc[/MENTION], or [MENTION=6785785]hawkeyefan[/MENTION] has suggested. So whose game, whose play, do you think you're pointing to here.
I'll start with 4e, because that's the version of D&D I know best. In 4e, there are three ways it can become the case that a PC can know something:
(1) The player imputes knowledge;
(2) The GM tells the player something that the PC knows, whether because of ingame situation (eg "You're in a windowless room") or because of background (eg "You remember that, as a child, all the householders in the village would sprinkle salt on the doorstep on the night of the full moon");
(3) The player succeeds at a knowledge check which obliges the GM to tell the player something that the PC knows (in some circumstances a successful ritual may augment or take the place of the check).
There is no way a player, in 4e, can establish a background element that
obliges the GM to tell the player stuff that the GM knows without requiring a knowledge check. That I have an adventuring uncle might be a bit of backstory that explains why I know about trolls. It might also explain why I have training in Dungeoneering. It can't
oblige the GM to tell about the weakness or immunities of Gricks if I don't already know, and I don't make a successful Dungeoneering check to gain knowledge about an Aberration.
I suspect that 5e defaults to much the same working as 4e, but will let [MENTION=6785785]hawkeyefan[/MENTION] elaborate if he cares to.
In AD&D or B/X the first two possibilities are there - player imputes knowledge; GM tells the player something - but the third option is much less common because there is no knowledge check system, just spellcasting. But just as in 4e and (I believe) 5e, so in AD&D and B/X there is no way that a player can establish a background that
obliges the GM to tell the player stuff, as PC knowledge, that the GM knows but the player doesn't.
In know version of D&D that I'm familiar with can a player, in virtue of a background about his/her adventuring uncle,
oblige the GM to provide information about the weaknesses of monsters that the PC encounters.
By bringing in player knowledge, it's no longer about PC vs. troll, it's player vs. game. It's what you know vs. what the game has in front of you.
If trying to metagame knowledge in signals to the DM that the players don't want to pretend not to know a weakness, then having an uncle tell you about the new monster signals the DM that the players do not want to lack knowledge of monster weaknesses.
First point:
it's not metagame knowledge if the PC also knows it. Which is what I, [MENTION=5142]Aldarc[/MENTION] and [MENTION=6785785]hawkeyefan[/MENTION] (at least - maybe other posters also) are positing.
You are asserting that the PC doesn't/can't know it, but no one else currently posting in this thread on this topic seems to agree with you.
Second point:
if a player signals to the GM that they don't want to pretend to be ignorant of a troll's weakness, why would you possibly infer from that that they want to be informed about weaknesses that they are ignorant of? To me, that doesn't seem a very good reading of human preferences, either in general or in the context of playing a RPG.