This has never been the problem in my mind. This is fairly self-evident. I'm simply wondering why this is considered a particularly salient point as a "strength" for the 4e rule set. You choose to play a paladin, and by golly, your character's mechanical crunch has them act like a paladin. This is hardly different than 3e, or 1e.
Are we simply assuming that "choosing to fight in a thematically appropriate way" makes it easier / more relevant to explore paladin-ey themes?
Not a problem for me either. My opinion is that it's achieved better in 4e mainly because of the clear lines between mechanics and thematics; the system openly acknowledges that these two are not the same, and that the latter can be changed arbitrarily as long as the former remains constant. This allows you to use the mechanical options presented as a tool for expressing a concept, rather than the more traditional view of a pre-defined package. You can think more in terms of "my character should be like xyz, what mechanical choices would let me play that?" instead of "my character is a Ranger, what can I do?" (though, of course, there is nothing to
stop you from doing the latter if you wish, which is why I consider this a better game).
Earlier editions - really, all RPGs that aren't rules-heavy and plenty that are as well - blur the line between what the game needs to function properly, and what the game designers thought was good flavour. As a result, they tend to rely on simply telling you how to describe a power or how to roleplay a class. In 3e/3.5, anything with an alignment restriction or variant makes a good example - they don't encourage you to show "suitable" behaviour by creating abilities than benefit from that behaviour, so much as they simply tell you "you can't do that" or punish you in-character for behaving in a different fashion. This makes them better defined, but also inelegant (because they tell rather than show) and rigid in terms of adaption to non-standard concepts.
I guess a lot of it is positive reinforcement versus punishment, and instruction versus encouragement. A 3.5 pally
tells you how to be a paladin, and punishes you if you don't observe its flavour correctly; a 4e pally
shows you how to be a paladin by giving you powers that encourage you to stand in the midst of foes and shield your allies, and gives you little return if you use those abilities in a sub-par fashion. Consequently, a 3.5 pally essentially refuses to be used for concepts that are at odds with its flavour, while a 4e pally happily adapts to anything that's not at odds with its mechanics.
In short, a 3.5 pally says "here is my flavour, play me right or choose something else", while the 4e pally says "these are my abilities, use them well or choose something else." Since using your abilities poorly will always be bad as long as there rules and tactical choices, I consider the 4e requirement to be inevitable, whereas I find the 3.5 requirement to be tyrannical and unnecessary.
3.5 pally is a paladin class you can play, while 4e pally is a class you can use to create a paladin. It's a subtle difference sometimes, but for me at least, it's a critical one.
If fluff is mutable, then by your own admission your objections to the barbarian class don't matter. The implied background of the barbarian doesn't matter.
Yeah, but
I didn't say "I don't want to play that because its fluff doesn't suit my character", I said "I don't want to play that because it doesn't do what my character should do". Things like wearing hide or plate armour, being a striker or a defender, or getting rage instead of channel divinity, are not fluff (though being divine or primal is admittedly on the fence).
You, conversely,
did treat flavour as law for case A, while totally ignoring it in case B. I'm pointing out that you're being hypocritical, not agreeing with your postulate.