OK, but I have to ask, why is Conceit wedded to emulation? Can I not explore a concept in terms of an entirely novel milieu or at least without regard to emulating anything?
How do you examine the Conceit without doing something to portray or instantiate it? "Emulation" is just the process of portrayal. As said elsewhere, it is the act of elevating a concept so that one might appreciate it. Sometimes this will mean Emulation in the sense of genre and the conventions so associated. Sometimes it will mean Emulation in the sense of a specific author's work (Jack Vance, HP Lovecraft, JRR Tolkien), others an overall theme or idea with no singular source that goes a bit beyond genre proper ("wacky hijinks," "survival," "intrigue"), or cultural packages associated with particular time periods and/or regions (e.g. "Arabian Nights," "Wuxia," "Sword-and-Sandal"), etc.
By elevating this tone, theme, genre, style, idea, etc., the act of play becomes focused on generating the conditions that will fulfill the thing elevated in a satisfying and effective way. Situations (challenges, issues, etc.) will be shaped not by their naturalistic-rationalistic consequences, nor by the need to surmount obstacles, nor by the values the players wish to see put to the test. (Or, at least, those things will be a secondary consideration, filigree overtop the main focus.) Instead, situations will be shaped so that the Conceit itself will remain center stage, and an enjoyable experience and exploration thereof can occur.
This is why you get DMs saying things like "the point of the game is FUN, if the system produces un-fun results then you SHOULD change them." Because they have an unstated commitment to a Conceit, action-adventure, which is more important than perfect Groundedness or (semi-)objective Score. (Generally most D&D folks don't engage overmuch with Values-and-Issues play at all, so that's neithet here nor there.) This then leads almost inexorably to fudging, among other tools of DM force like illusionism, to ensure the Conceit remains unbroken and center stage. But because most DMs also have a commitment to
either "the world is a 'real'/durable/tangible world, one that you can reason about and draw naturalistic conclusions from"
or to "there are real challenges in the world that, purely through your own skill, cunning, and resourcefulness, you can overcome and truly own the victory for doing so," the vast majority of DMs who use these tactics (particularly fudging and illusionism) intuitively know that they must do so secretly or it will "ruin" the experience.
It's also why the criticism of these behaviors tends to focus on either "well that's not very naturalistic, how can you say you value realism and physics-engine play and then turn around and secretly rewrite the world when you 'need' to in order to 'fix' it?" Or, from the opposite direction, "doesn't that invalidate the players' successes? They didn't earn anything, you just handed them victory." And then folks who say they stopped doing this and found great success are, essentially, saying that either pure(r) Score-and-Achievement play, or pure(r) Groundedness-and-Simulation play, proved successful and generated experiences the players cherished even though the Conceit of "high-action adventure" was
not always enjoyably front and center in the play experience (aka, Emulated). IOW, by this taxonomy, these are DMs saying they consciously chose to stop having C&E be the most important of their game-purposes, and instead allowed one of the others to be the most important instead.
Further, unlike Edwards, I have no issue with a game serving multiple purposes simultaneously. That the purposes are
incommensurate is perfectly cromulent with pursuing more than one. It just means that, in general, there will be one which takes greatest precedent (perhaps universally, perhaps contextually, e.g. S&A in combat and G&S in exploration). So D&D can be a game where the fundamental mechanics are geared toward defining how difficult obstacles are to overcome in a semi-objective fashion (Score) with player actions determining whether they are able to succeed in their goals or not (Achievement), while featuring various themes and tones as the central presentation of those adventures (Conceits) and extemporaneously portraying those themes etc. (Emulation), in a world meant to be modelled closely as if it were a real place with consistent physical-and-magical laws (Groundedness) that one can rationally predict and reason about (Simulation). Non-centric Conceit-and-Emulation play will mention or feature Conceits (as most games will), but will be willing to compromise on actually fulfilling that Conceit in order to maintain some other goal should the two conflict.
(As noted above, Values-and-Issues play is mostly vestigial in D&D, for a variety of reasons. So I didn't mention that, not because it can't be done--it can, as the "4e as Story Now" folks attest--but because that's just not typically a component of the...standard game-purposes "package" for D&D DMs.)
This has the added bonus of giving an additional explanations for why "incoherent" systems would tend to predominate even when many players have one particular preferred game-purpose. (Not that my taxonomy has "incoherent" as a thing, since I think that was a pejorative error on Edwards' part. I'm just responding to his use of it.) That is, people may genuinely actually want different purposes at different times within the same game, or may want to satiate multiple interests over the course of play, so long as each gets some attention. This is a direct challenge to Edwards' assertion (which may or may not have been explicitly stated?) that each person can value one and only one "creative agenda" in a given game. (I don't think he was so foolish as to assert that someone can't enjoy a Sim game for its Sim-ness and a Nar game for its Nar-ness. But he definitely seemed to have said "you can't enjoy both Nar AND Sim in the same game.")
I think it might depend on the TYPE of conceit!
I'm not sure what you mean. If the purpose is to enjoyably examine a thematic concept (whatever form that takes), and the system ever, for any reason, tells you that a result is inconsistent with that thematic concept, it seems to me you have only two choices: (1) accept the result because you value something else more than you value portraying the thematic concept, or (2) reject the result because you value faithfulness to the concept more than other things, and thus alter the result so that it suits. The secrecy of fudging and illusionism are a separate concern, namely preserving the false impression that the game remains fully G&S or S&A—aka "Simulationist" or "Gamist"—that arises from trying to, in essence, have one's cake and eat it too.
Do you have an example of a Conceit where this dichotomy (whether to follow Conceit or system when the two conflict) is not possible even in principle? Because if the Conceit is just "be like reality but augmented with specific fantasy elements," that isn't a Conceit anymore, it's just Groundedness (and this similarity is part of what allowed GNS to err in putting the two in a single bucket.)