I don't think I'm underestimating anything, other than humans' propensity for creating something out of nothing, though with no profit margin here, the market for this is small.
All I am saying is to clarify Endo's point, that essentially just about anything that you can come up with, will either be an aspect of an already existent path, or will consist of and overlap with multiple paths.
This is not necessarily a bad thing, but it does mean that you are limited in what you can accomplish here. You can add a Psionics path, but with current game mechanics, and all the spells that already exist, what will this path do? It seems most of what you'd want exists in Astral already, and while you could add a few more, the largest benefit would be in the creation of another bless opportunity.
This seems entirely out of the range of the modding community, only because the hardcoded structure of the game does not allow for it. If you are capped at 8 paths due to game mechanics that you can't change, then you have a bit of a conundrum. The first part being that you wouldn't be able to add blesses, not anytime too soon, but then you would also be limited to 8 of them regardless, so inventing new paths is somewhat silly, if all you can do is push around the spells that already exist.
If it were at all possible though, I think it's an intriguing idea. But broadening and diversifying the sources of magic would require a massive overhaul of the game, to the extent that you are envisioning. For example, if you have 12 paths, then likely the multiplier for pretender acquisition would need to be dropped from 8 to 6, as each path contains less power than previously. Likewise, all mages in the game would have to be modded, and many of them would require immense amounts of guesswork and testing to place them in the approximate power range that they occupy in the current game balance. Then you get the magic sites, and the gem issue would only exacerbate the problems with alchemy, where now you have 12 actual gem types, and so your chances of A) not finding sites, or B) ending up with gems you don't need, seem to both begin to rise. The obvious answer would be to make most sites cross-path, so that when you find one gem type, you usually find 2, however since you can only search for that site through one of the paths, your site acquisition will still be terribly hampered, you will essentially find 66% as many sites in any given game as you would now, because the others are in paths that your nation can't access.
The complications with this really are innumerable. It would be interesting to play with for another game, or really cool if they made the paths/schools fully moddable for the eventual Dom4 (people are acting like children lately, directly construing the "won't come soon" of Dom4, to a "won't come ever"), then if you wanted to really put in the effort, some truly unique and bizzare scenarios could be developed. But for now, there's just no way it can become anything other than an interesting mental exercise.