The perception of magic being an overpowering, one-solution-fits-all, kind of delimma led me to this idea. What would magic be like if it existed in some kind of quasi-space between reality and imagination? What if it could only effect you to the extent that you believed it could, but for those who believe the effects are truly real?

This is a very interesting thought, but I would advise against it. What if someone does not beleive in magic? It would not affect them then. So what you might end up with is a bunch of 'atheist' player characters... or NPC's. What if someone wanted to play a sorceror in a game that you are in and he would not be as effective as he usually would be? Would that player have any fun at all? I see this as a problem that will result in more problems later on in a game.
I do agree that magic has the propensity to be ultra-powerful, awesome, and lethal. However, it can also be lethal to its user. Most sorcery wielding characters tend to be reclusive in their use of their gift, and only use it in certain occasions where it warrants it, such as a war, a melee that a sorceror got trapped in.. etc. etc. etc.
So IMO, while sorcery as-is makes for unbalanced gameplay, the sorcery in itself becomes balanced by how you use the rules and how you weave the story-line. If a sorceror was able to use his magic openly and without fear of repurcussions beyond his physical self, I think it would take away some aspects of what sorcery is.. and it would be more common. Also, the Gift is very rare... and should be played as such. If there were alot of sorcerors in the world, we might end up seeing something like D&D's problematic system... and I shudder at that thought.
As for it being a one-solution fits-all kind of thing, I beleive it is not... unless that character has mastered all nine vagaries. Maximum starting vagaries is capped at 17. This is only 5 and 2/3rds out of 9... so you won't be seeing this sorceror solving every problem he comes across. And progression in vagaries is much much slower than that of proficiencies or statistics.... and might take even longer if a wise seneschal were to make the PC in question sit out for a length of time to research, study, and apprentice under a master.

-Ingenious

sounds like the mage dark ages style. I always liked the idea of belief effecting things. I fiddled with the idea of making the CTN higher based on the believability of the action. tossing a knight 50 feet up in the air is unbelievable. Having him fall on his face and knock himself silly, odd but plausible. The second would have a +1 or so while the first a +5 or so...
And if you made your magic affect the world in accordance with local myths legends and superstitions (banish the demon with silver a book and a bell etc) then the CTN drops....

Just an idea

This reminds me of a story told once by my seneschal and former DM in D&D. One of his friends, or his brother's friend kept saying 'I dont beleive that' whenever confronted with an illusion. I see the potential for the same problem here...
Your example of the knight being 'force-thrown' into the air 50 feet usually would be fatal for the knight... so he wouldn't be caring if he beleived that it happened or not afterwards.. he'd be dead. Witnesses would matter, but what if there were none at the time? What if someone who is caught using magic wipes their memories? This IS possible after-all in the normal system...

-Ingenious

Hello Ingenious, thanks for pointing out some of the pitfalls of this idea! I'm the sort that always thinks that every problem has a solution - it's all in the fine tuning - but you need to know the problems before you can start, so those points are really helpful. For example, the problem with characters all being non-believers could be handled by creating a mechanic for disbelieving. If characters really want to have a strong resistance, they need to earn it by perhaps taking a major gift and/or spending proficiency points, or something similar. I'm not sure how I would implement though, as this is still a very rough idea. Is a mechanic for belief even a good idea?

Ashren Va'Hale, I like your idea. That's a really great way to go for a "subtle magic" feel to a game, which I really like. Very cool. I think it would be good to quantify the penalty more concretely with a chart though in terms of in which ways something might appear unnatural and what types of laws of nature it breaks.

It's also different from what I was first thinking of. I was thinking of things actually appearing different to different people, which might bend logic to the breaking point. How do you explain multiple people experiencing the same event in different ways when you have the cold, hard, reality, of someone dying the true outcome? The knight being thrown 50' into the air is a good example. If the knight believes then he is thrown into the air and dies when he lands...but what do the non-believers see? Does he simply fall over? Do they see the broken bones and crumpled form or do they just see someone who died of fright? Maybe people tend to strangely ignore or gloss over these inconsistencies/evidence, but noticing them even indirectly corrupts their disbelief in magic such that it makes them harder to not believe in magic in the future...

What if everything you thought you saw in your own life was just smoke and mirrors covering up the unnatural magical events that happen around us - our rational minds disguising them in the same way that we re-construct the blind spot in our field of vision?

I'm not even sure if my idea is logically possible the way I originally envisioned it, but it seems like there might be a cool idea hidden in there somewhere.