I think we avoid giving it attention because it is, for lack of a better term, hard.
Yes. It's also at the same time, as most hard things actually are, intuitive and unreflected upon - like walking, talking, throwing and catching a ball, or how you know what a fellow simian is feeling.
But documenting how people actually play at a table? That's hard. From the very basic bits (you're playing an alter ego, it's a constructed imaginary world, etc.) to the idiosyncrasies (when is the player speaking in character and out of character, when can you "recall" an action .... aka, "I didn't mean to say that!" to what level of fourth-wall breaking is there allowed in the game?) to the ways that different rules are interpreted and used to the variances different tables will have in goals and what is "fun"....
I've on occasion thought to try to start the conversation by documenting the different undocumented processes that tables use to play the game. You mention one important one in your discussion:
"How does a table agree that a proposition has been made?"
It's a really important question that D&D has rarely addressed, and to my knowledge never addressed directly. Gygax seems to be addressing this question in at least some of the cases by suggesting you need a party "caller" or "leader" who validates that the propositions being offered aren't hypotheticals, but the one the party is actually going to undertake . But in his example of play, he shows that he also is willing to bypass the caller and directly interact with a player who is temporarily in the spot light, and willing to accept propositions that they player makes directly about their character. So there is some complex idea of group versus individual actions that probably underlie the declaration by the designer that the process of play requires a "caller", which not only most tables ignored, but which Gygax doesn't make explicit when he tells you to play that way (otherwise, it would have likely seemed more sensible).
Another related one is:
"What filter is used by the GM to determine whether a proposition, once made, is a valid proposition. And, as a side issue, how does the GM communicate back that the proposition is invalid."
An example of this might be the proposition: "I climb the wall.", where the GM decides that since there are several walls in the room, this proposition fails for being vague (establishing a metarule that propositions must be specific) and responds not with resolution of the request, but, "Which wall do you want to climb [and at what point are you climbing it?]?" This seems like a strange example, but it is entirely relevant to the very common proposition, "I search the room", a request that will be affirmed or be rejected depending on the table's proposition filter. Likewise, "I attempt to persuade the gaurd to let us pass.", is a request that may fail or may pass the table's proposition filter.
But most tables aren't even aware that they have a proposition filter, or that other tables may have different ones, and most GMs, if you point out that they've created a massive house metarule that governs their proposition filter and other GMs have massive house metarules of thier own, in my experience don't go, "That's interesting." They instead go, "Well, they are doing it wrong."
But what's even more interesting to me is that these metarules can functionally amount to two completely different games, even if both tables are playing with the same "rules".