I think I have to disagree with you again on the point "Mechanics in a game don't simulate any form of reality". 

It's my working assumption ( in my game world at least ) that game mechanics are meant to be a wildly simplified, consistent, simulation of an alternate reality - one where technology never progressed beyond early Renaissance levels, magic exists, higher beings/Gods intervene in the world, and Humans are not the only intelligent species. 

Yes - the rules are hyper-simple compared to ( any version ) of reality, but going down the slippery slope of detailed mechanics to describe every aspect of a pseudo-reality gets you 4th edition ( and that's fine if that's what you want 4th ed fans! I like parts of it myself ).

But I take the game mechanics - and those situations which fall outside the core rules - and match them in my head to what I think the world would most likely be as a result of the axioms I listed in the 2nd paragraph. Different RPG systems and settings have different axioms, and I judge their mechanics against their settings.

If this is not the case - if the rules are totally arbitrary, and exist merely to "keep things organized for the purposes of play" - then there's really no basic distinction between D&D and Chess or Go ( with purely abstract mathematical rules ), and you have no grounds for objecting to arbitrary rulings like "Only 1 elf allowed at the table", "Hexblade's can't use two-handers ever", or even "you can't do that, because I don't want to arbitrate that" - all the things being attributed to "immature GM'ing". While those are completely arbitrary decisions by the GM, if there's no "background reference reality" that doesn't matter as it's all arbitrary and those are choices  to "keep things organized for the purposes of play".

IMHO, A consistent set of rules is what separates D&D from playground "cops and robbers", and having a "base reality" is what separates D&D from Chess.

I do think you owe the players "this is how the world works". Part of RPG games is problem solving: here's a situation/conflict, here's an opponent, here's a goal, here's a set of abilities and tactics you can employ - how do you want to try and resolve this? If the player cannot count on consistency in your world, how can they tell what options are open to them, and how can they confidently predict what the results of their actions will be? Even if you have a "narrative justification" for the discrepancy in the rules from situation to situation, reasonably intelligent players will know that it's an "after the fact" justification, and even if they don't, you're eroding the players' sense of world consistency and confidence in their ability to predict the outcomes of the moves they attempt.

Player: "I'd like to throw a grapple over the branch and try and swing across the chasm like I did two sessions ago - does it still work like that?"

To me, it sounds like you're willing to sacrifice consistency, in order to avoid setting precedents you don't want to live with - and I can completely understand not wanting to get "stuck" with a bad precedent. The approach I take, is that I do make a ruling, I set a precedent, and until further notice, that's "how the world works". My players understand that's how their abilities work and they know how to apply them to the situation. However, I'm human - and sometime I set a bad precedent. In which case, I'll sit down with my players and say "OK guys, I screwed up. I know I said that you could do X, but it didn't consider the implications of Y & Z - so I'm not "ret-conning" anything, but I think I have to elaborate on the rule to deal with those implications, so moving forward, this is how it works ... is everyone OK with that?".

For example - I recently had a Monk join the party. They wanted to know how Stunning Strike would work. I ruled that they had to pre-declare using it before they rolled any dice, but if the attack roll missed, they didn't expend the Ki. In the back of my head, ( in my "background reality" ) this is the Monk "charging" their strike with energy,  but if they don't connect then it didn't "ground" and thus didn't discharge. Well ... damn ... what are the implications of this "energy model"? OK - well, I sat down with my Paladin, and explained why I was ruling this way with the Monk, and really, their Smite should work the same way - pre-declare it before you roll the attack ( charging the attack with "divine energy"), but only expend if you connect ( ground & discharge ), no more "OK, I hit? I wanna Smite now". Did I change my precedent? Yes - but I had an explanation based on "reality" - and in play I was consistent to old precedent ( and the player knew how it worked ) - and I am now consistent to the new precedent ( and the player knows how it works ). At no point did my players have any doubt as to how the rules worked, they could count on my game world being consistent to the rules in play at the time, and they could confidently predict the outcomes of the actions they attempted.

Again - this is only my interpretation - if what you do works for your group, then it's right for your group.

xguild:

I feel we're not really all that far apart in out positions, here.

My grappling hook example was overly broad - but it was meant to be an ( overly silly? ) example of how players need there to be consistency on some level - and I still believe that.

That level need not be on the level of the mechanics, however. In fact, you seem to be saying it can't be - and I can accept that.

Which is why I think one needs to fall back on the idea of a "pseudo-reality". There is a "higher truth" - the extrapolation of how the "real world" would work if magic existed, etc. , etc. - which the rules and mechanics attempt to approximate - and they might fail in certain applications and "corner cases" in which case the mechanics can ( and probably should ) change to match that case.

In my mind "how the world works" doesn't change - how the mechanics reflect that world is a bit more fluid.

As such - I don't think we're really arguing positions that are all that different.