What system are you talking about? 4e? 4e has no mechanic for turning the PCs into "minions" to fight much higher level antagonists.
Exactly.
Rather, it has a mechanic for turning those higher level antagonists into solos and the like.
4e thus gives formal terms to previously-informal variances within a group of monsters - you might have the 6 ordinary ogres with 40 h.p. each, the shaman ogre with 45 h.p. and spell use, and the chieftain with 80 h.p. that fought as a higher HD/higher level foe.
My argument is that any mechanic that turns the 40 h.p. ogres into 1 h.p. ogres is intentionally not being true to what's established in the fiction, and is thus very flawed.
This is because a game in which PCs are toggled either up or down would not make for very good play.
Agreed.
But as the PCs are a part of an internally-consistent (I hope!) setting, what applies to the PCs must then by extension apply to the rest of the game-world inhabitants; meaning that toggling them up and down is every bit as bad.
I've bolded a few bits which demonstrate that you don't understand how 4e's combat mechanics work. Because you talk about resolution processes as if they are part of the fiction. Whereas an obvious feature of 4e combat is that the resolution mechanics are not part of the fiction, and are not models of fictional processes, but are devices for establishing what occurs in the fiction.
I forget the exact term for it - dissociated something-or-other, it's been a while - but if the resolution processes and the fiction don't at least vaguely try to match up then the problem is with the processes, not the fiction.
This is actually true of Gygax's AD&D as well - Gygax makes the point that a single attack roll doesn't model one single bodily motion
Indeed; and here Gygax is in fact trying to match up the resolution process (one attack per round) to the fiction (1-minute-long rounds) by saying that the attack roll represents the best of many attempts over that one-minute span. In other words, he's taking a good approach.
but 4e takes this idea and develops it further.
Which, given that 4e rounds are but 6 seconds long, seems counterintuitive.
In Runequest a PC's hit points do not change significantly over time; but s/he typically improves his/her skill at parrying and/or dodging blows. It would be ridiculous to say that RQ is unrealistic because hp don't grow with experience and hence experienced adventurers are just as vulnerable to blows as inexperienced ones. Such a comment displays complete ignorance of the mechanical device that RQ uses to model increased fighting skill, which is not extra hp but rather is improved parry/dodge skill.
Actually I'd say that's every bit as realistic - and maybe even more so - than the non-4e D&D model.
Likewise and mutatis mutandis for 4e D&D. A paragon PC can kill a ghoul in 6 seconds. At the table the question of whether or not this take place is determined by making a single d20 roll and filtering that through the attack rules. If you narrate the fiction of that in a way which creates setting inconsistency then that's on you. It's not on the mechanics.
There's a word missing in the above which, if inserted, makes all the difference: A paragon PC can
maybe kill a ghoul in 6 seconds.
Look at it another way: unless you're fighting something that really only does have one hit point or less, such as a kitten or a small rat, there are three possible outcomes of any attack roll or sequence:
1. You do no damage at all (typically in D&D this means you miss outright unless some sort of DoaM mechanics are in play)
2. You cause damage to the foe but do not cause enough damage to kill* it
3. You cause enough damage to kill* the foe
* - or defeat, or subdue, or otherwise achieve your desired win condition.
Minion rules disallow #2 as an option, which is not only unrealistic but - again unless you're fighting a kitten - doesn't give the monster an even break.
A ghoul might normally have 30 h.p. and a paragon character might normally hit it for 4d6+20. Most of the time the paragon is going to one-shot it but there'll be the occasional time when she rolls really badly on those 4d6 and the ghoul survives with 1or 2 h.p. left - highly relevant if the ghoul then gets a good attack in and paralyses the paragon.
And a side note: this brings up another mechanic I've personally come to detest in all versions of D&D - all RPGs where it exists, come to that - and that's that, using the same example above, a hit can't do less than 24 points damage. There's a huge gulf between 0 damage (miss) and 24 or more (hit); and the greatest warrior in the world should still be capable of hitting for only 1 point damage on an unlucky shot no matter what bonuses she has going for her. To its credit 4e kinda waved at this problem a bit with some damage-on-a-miss mechanics, but to me a miss is a miss and thus 4e was coming at it from the wrong direction.
The far-from-perfect-but-better-than-nothing solution I use is that on any 'minimum' damage roll - here this would be 4 on the 4d6 - you add the bonuses to that roll (here giving 24) and then roll a die of that size to determine what damage you actually did. This means there's a small (sometimes very small, but never zero) chance that anything with more than 1 h.p. can survive a hit from pretty much anything - and the minion model again defeats this.