They are valid, but they require the GM to carefully craft the fictional situations so that they're present. This is much harder to do, especially consistently and for every fight, than just clear mechanical defeat condition outlined by the rules. That's why narrative defeat conditions cannot be the only defeat conditions for a combat heavy mass market game aimed at participants of varied level of skill and experience. This obviously does not mean narrative defeat conditions should not be present; they should and your game will be better with them.
Point of order: You have shown that the rules of the game cannot be written such that they expect only-and-exclusively narrative loss conditions. This is perfectly true and I have never said otherwise.
But there's an implication here: every game must have both mechanical loss conditions and narrative loss conditions, indeed, most if not all of the mechanical loss conditions that the rules support must always be present. E.g. "[narrative defeat conditions] should [be present] and your game will be better with them" only makes sense, in the context of the other things you've said, if you're already presuming that the bedrock of mechanical loss conditions is already present, otherwise there wouldn't be any loss conditions at all, which (AFAICT) no one here has ever asked for and which all of us agree would be bad.
I am only saying that, sometimes, for
some sessions/arcs/campaigns/groups, hard-requiring all of the harshest possible mechanical loss conditions is a bad fit, and likewise that for some cases, narrative loss conditions are better. I'd be a fool to dismiss all mechanical loss conditions, which is why I haven't done that. I'm just saying that the universal mandates claimed by some in this thread are wrong because of their claimed universality. Lanefan has explicitly said both that without mechanical loss conditions, all successes or achievements are "hollow" and empty; and that if even one kind of death is on the table, for some reason
all of them have to be on the table
all of the time. The only possible conclusion to draw from these is that anyone who isn't playing exactly as lethal as he does is doing it wrong, because their successes and achievements are "hollow" unless all types of character death are an everpresent risk.
Your preferences are possible and the new version of the game even added an optional defeated condition to replace death to specifically cater to such preferences. It still of course requires the GM to avoid framing situations where the defeated characters would logically die, and I think that limits quite a bit what sort of encounters and situations you can use, but I don't see a way around that limitation in any "no death" approach.
It doesn't require that. Instead, it requires either:
1. A modicum of preparation, as I explained earlier. I've almost certainly over-prepared for this, especially given the low lethality of my game, but it really doesn't take that much forethought. Having one ally NPC who would have a vested interest in their survival is all you really need.
2. Improvisational flexibility, where you set it as a mystery, you
don't know why this effect has occurred, but the group can work to learn why, how, and (oftentimes) from whom, and possibly learn to intentionally harness some of that power.
3. Systems already in place (as you noted already) that mitigate the issue directly, though I would try to avoid ham-fisted uses thereof. Subtlety is valuable.
You can still have situations where lethal danger is present. The only difference is that, barring intentional effort on the player's part (be that "no I want the character to die", "I accept the risk and try anyway", or intentionally going for "play stupid games, win stupid prizes" or whatever else), there will always be
some kind of way for the player to continue participating as that character,
if they wish to. There may be horrendous costs that the player is not willing to pay, or to ask others to pay; there may be horrible downstream effects from this event; there may be a lingering scar, debt, or sacrifice; etc., etc.
I make no promises that the world will be sunshine and roses, and in fact very nearly promise the reverse. The bright parts of
Jewel of the Desert are truly, sincerely bright; but the dark parts are truly, sincerely
dark. Chiaroscuro, not grimdark; the latter is as bad as Father Knows Best whitewashing, as both things result in being unable to see any contour or contrast. Screw up badly, at the wrong time or in the worst way? There will be permanent, possibily irreversible consequences.
I just do a little bit of effort to ensure that that one specific form of loss, the destruction of a character and thus all the emotional investment one has put into it, doesn't happen for light and transient causes. This doesn't require bending the world into a pretzel. As I have said elsewhere, multiple distinct players have, privately, thanked me for the robustness and consistency of the worldbuilding in
Jewel of the Desert. The seeds are there, waiting to germinate, and they
can be found and plucked out--if you can fit the pieces together early enough, which the PCs
have done. They've also failed to do it at times, and only realized the true meaning of something once the proverbial trap has been sprung. That rewards perceptiveness and creativity, without mandating that the PCs must succeed. Sometimes they do; I'd say often they do, because they put a lot of effort into preparing before they set out. But sometimes they don't.