Thanks for the many compliments and XP rewards, folks. I'm actually a little surprised how well my idea was taken given the way many people feel about alignments. Glad to see that a lot of people liked it.
I think that there should be no required mechanical interactions with alignment. With respect to the Detect Spells I can think of 3 distinct options to define how they relate to alignment, which might serve in most campaigns.
1) They don't exist. Done and done.
2) Traditional, detecting evil creatures and effects.
3) Compromise, detecting only current evil actions, states, or thoughts.
Detect Alignment Spells and other mechanical effects of alignment should be an option, I think, but an option. As in, players and DMs are encouraged to take it or leave it based on their own preferences.
While your idea of a compromise is interesting, it seems a bit obtuse. Is it just a snapshot of a single instant of all nearby creatures? Is it active all the time?
Still, I can see one really good aspect to the idea: it keeps things ambiguous. Detect Evil doesn't just become a cheat-sheet for figuring out if an NPC is a legitimate target or not, it only tells you the current intent of a creature and can misfire, detecting good creatures as evil. But it sounds a little complicated.
[Nivenus' proposal] basically says, "You can be dedicated to Good, Evil, Chaos, Law, with the possibility of having multiple allegiances. If neither of these are satisfactory, you can declare yourself True Neutral or simply Unaligned."
That's basically the idea. You also got at another point, which I think is worth considering, which is the idea of five (or four) core alignments, with the rest being considered secondary hybrids (an idea I posted earlier today on the WotC forums).
Basically, you could keep the five-alignment system from 4e with some modification to make it closer to the original alignments from 1e. You have Law, Good, Chaos, and Evil, along with Unaligned as the core alignments from which all the others stem. For people who just want a clear-cut character archetype and don't want to worry about middling little details, they're the easiest way to go.
Then you have the secondary alignments, Lawful Good, Chaotic Good, Lawful Evil, Chaotic Evil, and (True) Neutral, which represent characters who have a more complicated moral and ethical outlook. As a result, you basically have both 4e and 3e's alignment system in one, without significantly compromising either.
It could devolve into nothing more than uncaring vice unaligned (which puts it closer to the intersection of weak Chaotic and stronger True Neutral, in my head) I think there's enough headspace to differentiate between TN and UA if we define True Neutral as strictly those who actively resist the Four Axis construct (Law/Chaos/Good/Evil). Unaligned wouldn't have that active component.
The way I see it, True Neutral characters believe every action has an equal and opposite reaction and that doing little to impact the world results in the least possible suffering. Unlike the silly idea of druids who backstab their allies when the going gets good, this one has an actual basis in moral philosophy - Daoists to a certain extent share this view through the idea of "wu wei" (action without action).
It would be a very difficult alignment to play - True Neutral characters aren't going to rush into doing anything they think might upset the cosmic balance or cause unintended harm, so they're not going to be the most impulsive or active heroes. But it
is playable, so long as the player has the right mindset and it's the right kind of game, and it should be available.
Again, though, it and the hybrid alignments might work best as an "advanced pack" of alignments, supplementing the core 5.
I remember 5 alignment, remember when dual alignments didn't exist? Just good, evil, N, lawful, and chaotic?
One of the reasons I'd propose splitting the alignments into two categories: core (pure alignments) and secondary (the dual alignments and True Neutral).
On another note, I think it would be useful for each axis of alignment to cover a different kind of behavior. Good and evil, it seems to me, should be more about how you treat people on an individual basis, while law and chaos is more about how you relate to society.
"Greater good" questions, for example, which I think are often the root of alignment arguments, should be the purview of law and chaos, not good and evil.