Ok, giving you the benefit of the doubt, you asserted: "From a player perspective, I’m discovering what’s in a dungeon as I make my way through it, whether it’s generated ahead of time by the GM or whether it’s generated though some procedure of play."
So back up in post #194 (
GMs: Guiding Morals in GMing) I wrote:
"Well, that "generated through some procedure of play" is pretty darn broad. I concede that there are some procedures of play that are in fact equal to being pregenerated. An example of this would be purely "procedural generation" in a game like Nethack..."
I then went on to explain why, even though it was theoretically possible that some continuous procedure of play could in theory replicate the experience of pregenerated myth, it was highly unlikely in practice.
Then, after I had explained that, you then wrote in response to my post: "So what if I use multiple tables to generate a dungeon and its inhabitants and their possessions? If I use this method a week before play and write down the results, then that allows for discovery… but if I follow the same method at the time of play, that doesn’t?"
Now, that's baffling, because if you'd actually read my post you would have known the answer to your question. If you use multiple tables to generate a dungeon through a strictly procedurally process, then - as I already explained- it doesn't matter whether you generate the content ahead of time or at play time because in both cases you the GM have reduced your ability to choose down to zero. In fact, I spelled this out as specific example of a procedure of play that was theoretically equal to pregenerated myth before you first raised the question. You can go back to the above post and read my discussion of it.
In the larger context, what the GM has done here is replace some of the shackles on GM agency in the form of established myth with one of the other shackles on GM agency - fortune - so that ultimately the GM is still reducing their ability as a GM to influence the setting according to their own wishes and thus giving space for the player to have agency. And in both cases, you've also provided for Discovery because the players doing the discovering still have something to discover that is real, which is in this case the intricate tables and processes you are using to simulate human creativity and the concrete setting that those tables ultimately produce.
Now as a practical matter, as someone who has used random tables as a prompt to the imagination, I think you'll find if you try to do this as a continuous process with something like generating the room after the door is opened or the corridor is traversed, it doesn't actually work to replicate pregenerated myth because you end up generating content which should have influenced prior rooms. Content "leaks" between locations through things like sound, water flows, scent, light, and influence (monsters living in "A" probably know something about what is in "B" and act accordingly). And because this random process is more suited to a computer than a human rolling a dice, you find that you can't do it seamlessly unless you pregenerate sufficiently large sections (as for example Nethack does by generating whole levels at a time).