But isn't the idea of heroes and superheroes already based to some extent on these same literary and historical exemplars? I find the examples of Druid and Monk just further my argument that the exemplar is key in the decision to frame the idea of classes at all. If I want Druids in my game, as I see in history or literature, I need (or perhaps really want) a way to represent Druids that's different from the basic Hero. How do I do that? Partially with evocative fluff, but more importantly with a bespoke set of skills and abilities. So yes, that indexes game-piece distinction, but it also strongly indexes the exemplar being differentiated.I think I'm not as sure that the original game pieces needed a particular trope or exemplar to work. Heroes and Superheroes can be Aragorn, Eomer, Conan, Lancelot, etc without worrying too much about details beyond that. And all we need to know about Wizards, exemplar wise, is that they use magic. The artillery-like effects are a product of wargaming requirements and conventions, plus maybe a dash of Dr Strange.
The role of classes, and their relationship to tropes, has changed a lot since Chainmail; and beginnings of that change can be seen even with the Druid and the Monk. I still think that, in the original conception of the game, game function was primary.
The early wizard doesn't change this equation much either, it is present specifically because of a desire to have those characters (like Merlin, or whomever) in the game next Enkil and Hercules. The access to magical powers is the obvious mechanical design difference. But again, without the exemplars, there'd be no wizard. I think it's really, really tough to pull D&D apart at this historical moment to separate game piece from literary/historical exemplar. You simply don't get D&D without both working at the same time.
I suppose that all these powers could have just been heaped on the hero and added via point buy as needed, but there we are already dropping a great whopping pile of anachronisms into the discussion. I think the explanation that makes the most sense while assuming the least is that Gary and Dave saw Wizards and Heroes as different things and so designed them as different things represented by different mechanics. On from there we eventually get to here.
The extent to which framing is handled well by game A is all over the map. It's certainly not a historical march toward best practices. Games continue to do this well and poorly. My opinion is that this is a function of how well or poorly the designer in questions actually understands what's happening at the table, as opposed to what's happening in a rulebook.What I'm sensitive to - most recently, I was commenting on it in relation to Mythic Bastionland - is when a RPG doesn't tell the GM how they should be making framing decisions.
Classic D&D does (although it doesn't use the language of framing) - the GM frames by reference to the location of the PCs on the map, with reference to the key; and also by reference to the outcomes of wandering monster rolls. Burning Wheel does too, albeit it's rules for framing are completely different from classic D&D. Apocalypse World is a little bit more indirect, but does set out a method for framing - the interplay between the prep of fronts, and the making of GM moves.
But the GM-as-world-mediator approach, at least when explained by its proponents, tends to eschew giving an account of how the GM is to make these framing decisions. It doesn't want them to be done by reference to rising action, or theme; but what approach should be used is not normally explained.
In the specific case you mention the most common reply I've seen is that the GM adjudicates consequences based on character actions and changes the setting appropriately. Obviously, this sounds like framing without calling it framing and has been something I've argued with folks about quite a number of times. I suspect that it's up to the individual DM and their prior experience how well or poorly this is accomplished.

