Yeah, well, the problem comes when the DM's monsters know that your Fighter has Combat Challenge and Combat Superiority, but you don't know when you're marked that the monster marking you will smack you upside the head for 2d8+x (or whatever their damage is) if you ignore it. But, that's more of a DMing issue, and not a true rules issue.
Right: the DM has to be good enough to separate "What this monster knows" from "What I know".
But that's a given for the game, in general.
The same is true for just about anything role-playing-wise: the player has to be able to identify what he knows but his character doesn't, the DM has to provide means of getting information to the player that his character knows but the player doesn't.
Thus:
You DO have the chance to know.
That's fully covered under monster knowledge checks, under 'powers'.
The rules are actually pretty blunt on that issue.
Which makes me think that a similar, inverse check would be required for the monster to recognize anything beyond the basic "I'm marked" condition.
I don't think there's anything that supports a strong argument either by RAW or by RAI to settle the issue "Do you know everything that can trigger by ignoring a mark, or do you only know that the marked condition has been applied to you?"
. .
The argument "if any Level 1 character can ..." is a bit of a straw man, because it assumes Level 1 player characters are very common. By RAW, the player characters are actually
unique: an NPC of the same class isn't built by the same rules, so the chances of the monster having encountered an actual PC before? Probably fairly rare, unless the DM thinks "Yes, Level 1 player characters are common in my world."
Then there's also a question of how much interaction there are between the races: if you're DM'ing a world where Eladrin are considered a myth, a forgotten race from an earlier time .. but somebody's playing an Eladrin, well, monsters should really have no idea what he's capable of. Even if the Eladrin are simply sequestered, typically keeping to themselves, the typical orc may not know much about them .. while if there's a long history of warfare between orcs and Eladrin, then, yeah, pretty much every orc should know about the whole teleport-y thing.
The same question applies for your Swordmage example. Yes, every swordmage has an Aegis. But if there are only fifty living Swordmages, it may not be likely that every intelligent monster out there can recognize one right off the bat.
Addressing class-specific marks: if I'm wearing chain mail, long sword, and small shield .. what class am I? I could be just about anything - I've seen a paladin in that get-up, as well as warlord, fighter, and a few others. The idea that the monsters automatically recognize the
class of the enemy, and therefore all of their powers, feels a bit .. off .. to me.
Which may be why the RAW is intentionally vague on this issue, as discussed by Gruns and I on page 1 of this thread.
And why I'll play marks much more the way Prestidigitalis seems to want: some monsters may recognize things pretty quickly, some monsters will learn from experience, and some monsters will be too dumb to ever learn.
Really, that's no different than other powers: some monsters know that magic is more dangerous if they cluster together, some monsters need to learn that from experience, and some monsters will be too dumb to ever learn.
If all monsters refuse to cluster together, that really sucks for the arcane casters, your Wizards, Warlocks, and Sorcerers especially. If all monsters always know everything that a player can do on a mark, that really sucks for the defenders.