You confess confusion near the end of your post, and I think it is because you are thinking I disagree with you far more than I do.
But, I think it is rather inappropriate to label either of our viewpoints with negative terms like "cynical".
I call it cynical when someone says that they want a new edition of D&D but think that the Fighter class (the Fighter class!) needs to be kicked to the curb in favor of well, something that sounds like an almost unplayable and unbalancable mess of self-contridictory viewpoints.
I'm actually rather optimistic about 5E. I vastly prefer to like things rather than hate them, and my wishes are closer to being a hope for positive change rather than a despair about an unfixable solution.
I'm actually rather optimistic about 5E as well. And in no small part because I suspect or at least hope that Monte Cook isn't going to do the sort of 4E and Pathfinder inspired stuff that you are drawing your inspiration from.
The real issue, I think, is that I simply have very different likes and dislikes than you.
Perhaps.
I have different ideas for what a well-designed class is than you.
Certainly.
I want different things from the game than you.
Probably not. I suspect from what you say that we want much the same things, we just have very different ideas of how to get there.
There are a lot of reasons why they can't (or rather, shouldn't) be done in a single Fighter class.
Feel free then to explain them.
A big part of this is that I don't hold a single Fighter class as an ideal worth striving for.
Of course not. If it could be done better with lots of class variations, then I'd support it. Problem is, we have examples of that in 3.5e, 4e and Pathfinder and it made things worse and not better.
In general, I prefer a larger number of mechanically simple classes with clear focuses, rather than a small number of 'toolkit' classes designed to be customized into other things. I think my preferred approach has more mechanical elegance and brings out the flavor of a class better.
Large numbers of classes are elegant? So, you consider late 3.5 to be mechanically elegant?
I mean, let's look at the example of two reasonable non-magical combatant archetypes: the heavily armored melee fighter and the lightly armored archer. These two archetypes are almost complete opposites, with opposing strengths and weaknesses. The armored knight would be able to safely take many hits and is skilled at fighting in close quarters, but would be very ineffective at ranged combat. On the other hand, the archer would want to avoid close combat at any costs, but is very good at fighting from a distance. In many ways, the archer is more similar to the stereotypical Wizard than he would be to the armored knight, and the armored knight is more similar to the Paladin or Barbarian than he would be to the archer. Those two archetypes belong in different classes, not the same class.
There are all sorts of bundles of assumptions in that which we first have to dismantle before we can discuss why this approach won't work. Why for example should a wizard be focused on ranged combat? Shouldn't a wizard also be able to focus on close combat, or on some thing entirely different than being 'artillery'? I think you'll find plenty of people upset with 4e pigeon holing wizards into the 'ranged support' role in the first place, and rightly so. Must we know have separate classes for wizards up close, and wizards at a distance, and wizards that do different things entirely? And perhaps some wizard in the middle so that if you want to play something more balanced and less hyperspecialized, you can play that as well? And for that matter, why can't the Paladin or the Barbarian be ranged combatants? Do we need mini-classes for the elven Paladin that hits hard with bows? For the Barbarian archer, or rock throwing warrior based off of a Polynessian archetype? Or for that matter, why should we assume 'Viking beserker' is archetype Barbarian? What about native American aboriginals mounted on horseback? Not barbarians? And for that matter, why should we assume that all warriors who channel their rage and emotion into ferocity are rural illiterates from more primative cultures? Why not elite nationalist units, or sworn temple gaurdians? Do we need separate mini classes with their own subsystems for those concepts as welll?
So, what about gladiators, axe wielders, lancers, crossbowmen, flail specialists, whip wielders, wrestlers, martial artists, blowgun wielders, boxers, knife fighters, peltasts, mounted archers, pirates, halbred wielders, fencers, sword and board, legionares, slingers, hoplites, pirates, muskateers, men-at-arms, ninjas, samuarii, knights, florentine style fencers, sumo wrestlers, charioteers, and so on and so forth. Should we have a mini-class for each concept and style? And what happens when my character concept wants to borrow from multiple classes but is forced to pick up unwanted baggage from each? What about smart fighters, cunning fighters, charismatic fighters, agile fighters, tough fighters, and on and on. Can you support each approach for all the little subclasses? I mean, surely you aren't suggesting that every knight needs to take the same basic array of attributes to play the class? What if your powers don't interact with the attributes my concept calls for? Must I be suboptimal?
Up front it takes much less effort than creating a unified class. The problem with a new class for each problem is that its lazy design. It's like solving a coding problem with a new batch of in line in place fall through code. It's simple in the short run and fixes a specific problem, but in the end it leaves you with an inelegant mess because you only fixed that one problem.
Their differences outweigh their similarities, so trying to create a catch-all Fighter class simply serves to diminish your ability to create any of them.
No, it doesn't. It only does so in your experience because the attempts to solve the problem by Pazio and WotC and others have been poorly implemented. Take the example of the archer versus heavy armor close combat specialist you are focused on. A lot of the distinction between the two can be defined very simply - one is focused on traditional dexterity roles and the other is focused on traditional constitution roles. One is mobile by virtue of light armor, and the other is hard to take down by virtue of heavy armor. We have major distinctions between the two using nothing more complex than what was available in 1e. There are plenty of tools for providing for extremely high distinction between the two even if they have a unified class.
I'll also admit that I'm not at all a fan of a 3E-stlye skill system. I'd rather see such a system removed entirely, rather than elaborated upon. I think that'd be the subject of a different thread, though.
Ok, sure. But i'd like to point out that skill systems are a lot easier to remove from the system than they are to put back in when missing, and that D&D has had skill systems since late 1e at the latest and much earlier if you count the thief as the prototype upon which the notion of skill in the game is founded. I understand why the OD&D players dislike the thief, but if you are going to reach back that far for your inspiration on how in game skill should be handled but on the other hand plan on having literally hundreds of classes in the editions long run, then I think you are imagining a game that is unlikely to unify the fan base. More people left OD&D for its lack of skills than stuck with it for its lack.
That doesn't mean non-magical should equate to "basic attacks and skill checks" while magical equates to "complicated sub-systems".
Did I suggest that it should? Tactical subsystems can be quite complex without depending on the sort of limited access balancing or point expenditure balancing that is associated with powers and manuevers. And limited access balancing which is grounded primarily in metagame needs is basicly a magic subsystem. In particular, one of the problems with associating maneuvers with a class or power, is that if this subsystem covers a set of actions that are not meant to be supernatural, there is a question of why access to the subsytem is restricted. For example, if I have manuever that is 'Judo throw foe', the question becomes "Why can't anyone pick up something and throw it? Sure, maybe he's better at it, but any 5 year old can body slam something his size." "Why can't I pick up a handful of sand and toss it in the foes eyes? Sure, he may be better at it, but can't I try?" You'll run into the problem of, "Well, you can't do that because its not a power on your class list." really really quickly if you try to address the problem through numerous classes and subsystems.
Personally, I think there should be a wide variety of very different sub-systems of varying complexity and style. Non-magical warriors need to be able to do cool stuff, and things like maneuvers and powers are perfectly appropriate for that in my book.
Powers are spells by a different name. While I'm inclined to agree with you about the utility of separate subsystems to a certain extent, at some point power subsystems are just generic magic subsystems in thin disguise.
A Cavalier class would be useless in a dungeon-focused campaign.
That said, the mounted cavalry soldier is one of the more absolutely important warrior archetypes in history. For any campaign that focuses on something outside of the dungeon, mounted combat becomes really, really important and interesting. A Cavalier class would be really useful in any of these campaigns.
Again, we disagree not about the end goal, but about the utility of having a Cavalier class be the means to get there. Let me put it this way, do you agree that a mounted warrior that rides a horse, and one that rides a gryphon, and one that rides a shark, can be bundled into the same class? Or would you prefer separate subsystems for land based mounts, air based mounts, and sea based mounts? And, if you do prefer that, then what are you going to do when those subsystems are tied to a class (rather than to the environment) and a member of another class wants to interface with them?
Generally speaking, I'm not a fan of the idea that any class, class feature, or mechanic needs to be equally useful in every campaign.
I never said it needs to be. But I'm equally not a fan of creating a class specificly because some campaign might occur entirely in the ocean.