If the players have no clue how the world works, it must be a free-form game with a GM that's told them nothing about it. Even rules light games give one a pretty good idea how the world works. The "meaningful decision" argument is a good one theoretically, but I've not come across it in any game I've played.
Having "no clue" of course requires "no rules" and, as you say yourself, this is generally not the case even at game start. Lack of rules still leads to lack of knowledge, however - and I have come accross it quite a bit. Usually, it happens when players try to get "creative" with their plans, and it manifests as the players believing that a plan is realistic (in the sense that it has a good chance to work, more than that it would work out the way described in the real world - but both are factors) when the GM has (effectively) decided that it's a dud.
The problem, from my point of view, for rules heavy games is that they aren't fun for me to GM. Two things about rules heavy: 1) more rules rewards rules memorization and mastery. So, no need to be creative when it comes to interacting with the mechanics, instead I get to use rote memorization. Fun.
With rules that are just "weighty" that's an issue, I agree. Which is why I think the epitome of "good rules" aims for
elegance - which is covering lots of ground with short, clear and eloquent rules. I think 4E does this well, but I would also recommend Primetime Adventures as a (totally different) game that does it exceptionally well, too. Dungeon world seems to have some merit, judging by what I read of it, but I have yet to sit down and read it (I only just got it).
2) It implies that the GM can't be trusted to be fair, so as more rules are added, more GM neutrality is a given. GMs should be impartial, but that shouldn't require walls. If the players don't trust the GM to run a good game, there's problems beyond the rules.
For some it may be a trust thing, but I don't think that's the core of it. It's a world-model thing. Some people just have beliefs about how stuff works that are radically different to the way I think stuff works. That doesn't mean either of us is wrong (though, obviously, I think they are, otherwise my beliefs would be different!), it just means that what each of us thinks will be the result of a situation will differ - which is a problem if one of us gets to dictate what happens without the other of us knowing what the "model" is in advance. Rules avoid this by encoding the model in a set of dice relations (or whatever) that we can both read and understand. The fact that, for example, you think that raising your greataxe high above your head prior to a strike will increase your damage while I think it will just give an opening for your opponent to kill you will not matter if the rules state (in some way) what the outcome of such a move will be. Often, I find it works best to say that the move makes no difference to the standard combat system process, and just allow players to describe/explain/rationalise the outcomes however they wish. Otherwise Thrud the Barbarian would be a dead bunny in all my games...
So, as we approach more rules from less rules, rules mastery and impartiality are rewarded. So, what's the pinnacle GM in regards to rules mastery and impartiality? A computer! So, basically, when running a rules heavy game, when I interact with the mechanics, I should do my best to emulate an AI. If that's the case, I'd rather an actual AI take the job. Being the intermediary between the book and players isn't my idea of a good time. Making decisions, as opposed to rules regurgitation, is the best part of GMing, if you ask me. Mileage may vary on this front, of course.
Rules application the computer does very well indeed, but long experience tells me that tactical play of combat opponents, character play of NPCs and strategic planning are things the "AI" does exceedingly poorly most of the time. My idea of the ideal game is a computer (or equivalent through steamlined, elegant rules known to all the players) handling the interactions in detailed time and a human (or near-human) GM (or other players) handling the plotting and planning.
All that said, that's just an argument against the chunky rules system, not the abolishment of all rules. It's not a game without rules.
Right - hence the absence of "total cluelessness", earlier. But the players are clueless to the extent that the rules don't cover the interaction between their character and the rest of the world. Hence it should be covered as far as possible. The rules don't need to be "heavy" - this is not a plea for rules that are "realistic" or cover every possible factor that someone imagines will be pertinent in "real life". The rules just need to be rules - not vague models that get overruled the instant some individual thinks that they "obviously don't apply". "Obviously don't apply" is in the eye of the beholder, and if I'm relying on the rule being a rule for my character's action it sucks to find out that the (game) world actually works differently than a person who has grown up there thinks it does.