As I said, NPCs used PC rules to determine their abilities, except in situations where the in-game reality diverged. An orcish shaman might not have the same spell access as a PC priest, but only in as much as a shaman is actually different from a priest within the world.
My take on that is they are only different in as much as there are different traditions within humanoid tribes. The shaman/witch doctor rules reflect an easy way to deal with that which is far simpler than making up some sort of 'NPC Class' that adds lots of extraneous detail. Any given specific shaman might well have different spells than those listed, possibly reflecting some individual idiosyncrasy. 4e, and 5e AFAICT instead prefer to just stick to the stat block approach where each unique individual creature has its own stat block which might potentially contain anything. Of course 4e has a 'goblin hexer' stat block that represents a pretty typical shaman of Maglubiyet. I suspect the 5e MM has a similar one. In 4e at least, you might present said character as a minion at high levels, though truthfully its unlikely to be worthwhile.
As I said, the game mechanics reflect the in-game reality. Ogres could have better AC by wearing better armor, because that's how armor works within the world, as confirmed by the rules for when a PC wears armor.
The 4e monster rules also provide for this, but in general the advice is to level up monsters. The theory being that all the stats on monsters are generalizations, and better armored ogres are no doubt also better trained, equipped lead, motivated, etc and thus worth being considered higher level than unarmored ogres. The increase in AC is 'caused by wearing heavier armor', but its all part of a coherent whole. You wouldn't describe a level 15 ogre and a level 9 ogre in the same way.
Until 4E came around, it wasn't something worth talking about, because it was just one of those obvious things that didn't need to be said. It wasn't a particular principle of D&D, because it was such a fundamental assumption of all RPGs that there was never any reason to question it. Kind of like how gravity affects everything, all the time, and you don't usually need to mention it unless it changes suddenly.
Again, I strongly reject that there is any principle that game rules correspond to 'reality' in the games in question. This is clearly not the case in many games, even early ones, such as T&T, or maybe a famous example would be 'Toon' which doesn't HAVE rules for its game world at all. It just wasn't a question in D&D, as the style of game is GM-Centered.
Hit Points don't exist, but the realities of the game world which are reflected in the Hit Point mechanic do exist, objectively. It is objectively true that a given character possesses certain characteristics of luck, skill, toughness, and whatever else. It is a true fact of the game world that it takes an average of X number of 'hits' from Y weapon before a given character will be unable to continue fighting, for whatever definition of 'hit' you choose to employ. Or to use a less abstract example, it is a true fact of the world that some people can survive a fall from any height without dying. Hit Points reflect a real, objectively measurable phenomenon.
No, not really. A given creature for instance might be much less capable against some opponents because its divine favor doesn't work against them (maybe they're more favored by the same god). This could for instance explain how wizards tame creatures that could easily kill them by simple mechanical extrapolation. Nor is it necessarily true that people can survive certain falls. PCs survive certain falls because the DM uses consistent rules to cover them, due to the fact that the players will feel helpless if he makes arbitrary rulings (amongst other reasons).
Mostly though, we don't know the recipe for hit points, so all those factors that go into them, we don't actually have any way to know what they are, or if 2 creatures have the same hit points for the same reason, or even if the same creature's hit points always represent the same thing. A GM would be perfectly free to say that a given creature died because its hit points are all luck and its not its lucky day today, so a fall killed it, and another creature survived the same fall with the same damage and hit points because all its hit points come from toughness.
Definitions are useful in as far as they allow meaningful discussion and inform decision-making. I consider the objective reality - the consistent-stat-representation-requirement - to be such a fundamental rule of RPGs that I would not consider games which violate that rule to still be in the same category of game. If someone refers to something as an RPG, then that word carries a lot of meaning to me, and I would be disappointed to later find out that it was this other thing - that it didn't follow the basic rules for what makes an RPG.
Yes, definitions are crucial for communications, and when you use one so idiosyncratic that I doubt any other poster on this or any other forum I regularly post on would agree with yours, then you will probably have communications problems.
I would also point out that your definition is IMNSHO political, not rhetorical. You don't like 4e so you literally want to define it out of existence. This is a political act, not any sort of an attempt to communicate clearly, and trying to plead otherwise is disingenuous.
You can shift definitions around in any way that promotes decision-making and allows for discussion. Some people don't consider StarCraft or Magic to be real sports. Some people don't consider RPGs to be real games. By sticking with my definition, I'm trying to draw awareness to this phenomenon, and to get people to understand just how big of a deal it actually is. This is a huge deal. It's not something that should be changed without significant consideration. Breaking this rule will turn away a significant portion of the player base.
Well, again, you're betraying the purely political nature of your 'definition'. Nobody else agrees with it, its offensive to some, and you'll most certainly find the grade of discussion you will have when you invoke said nonsense to be lower than if you don't.