D&D's Barbarian has nothing to do with being a non-Greek, non-Roman, or Non-Christian.
D&D's Bard and Druid has nothing to Medieval Celtic Priesthoods.
D&D's Cleric isn't Muslim in faith.
D&D's Monk isn't even Religious!
D&D's Paladin isn't a Knight of Charlemagne, etc.
Similar problems arise with Ranger, Sorcerer, and Warlord: they define a game-role with a term or profession that has connotations beyond D&D.
With all due respect...
A cleric, for example, is "a man in a religious order, a man in holy orders". There is no conflict of meaning here.
Yes, the cleric, barbarian, bard, druid, monk, paladin -- these have real-life historical and/or modern context, but they also have a commonly understood generic meaning. Sci-fi and fantasy is replete with these archetypes appropriated into an imaginary setting.
Thanks to Tolkien, rangers are so common in fantasy that everyone "gets" that too. Just as importantly, the fluff for rangers in RPG is more-or-less believable. That's what good fantasy does: translocating real-life concepts into fantasy settings in a believable satisfying way. Furthermore, there are no other types of rangers in D&D (no Texas rangers, no government protected park wardens) so there's no paradox.
"Sorcerer", "wizard", "warlock" in real-life are used so vaguely and interchangeably that there is no authoritative definitions and fantasy writers can feel free to define them as they wish.
I agree with you, so why is Warlord the one exception?
For the same reason that "bloodied" metagame does not mean "bloodied" ingame (see post #977).
A Warlord is not necessarily a warlord, and a warlord is not necessarily a Warlord.
(Also see post #850, 865, 876, 898, 901, 916, etc.)