You could make it tied to subclass on classes that get them at 1st level. You could make a feat to switch the dedicated stat.
You could make everything more mad, but why would you want to?
I don't see how making a caster need multiple stats for the same feature is any different than making weapons need one stat to hit and another for damage.
Instead of making a feature worse by making it more MAD, why not incentivize using the other stats? Figure out or create a reason why each spellcaster would want to specialise in a stat other than their primary or constitution.
You'd probably need a different reason for each class to want to use each secondary stat, which is a heap of work.
Whichever way you do it the result will be the same. A Ranger needs Dex and Wis, and Con like everyone else, plus they need Strength if they want to go Melee. If you want to make Int appealing as well, then to be worth dropping something else, they need it has to be
really appealing or else it's just a consolation prize for being sub-optimal, which of course means that in most cases people will still dump it, but just be more annoyed about it.
And if you're offering strong incentives then you might as well do it in the simplest way possible.
In theory Str for Damage and Dex to hit is fine. There's a trade-off between hitting more and doing less damage and hitting less but doing more damage. In practice, it's more difficult. Part of this is because, even when the trade off is even, it's better to hit reliably then be swingy, and part of it's the psychology (people don't like missing). However, this can be dealt with by actually weighting Str a little more in the trade-off. The real problem is scaling. Playtesting has to be rigorous because it's very easy for a source of bonus damage to swing things towards Dex (The less percentage of your damage that comes from Strength the less valuable it is). WotC learnt during the rushed production of 4e (and the errata they produced for it over the years) that the easiest way to balance is to simply not have the kinds of interactinos in the system that require rigorous playtesting.
The trade-off that 4e brought about between Str and Dex in that Dex became irrelevant to heavy armour uses completely. Pre 3E Dex was always good for AC no matter what armour you wore. In 3E you could do this too but it was never quite as clean, as you often wanted some Dex to meet prerequisites, and each type of armour had a different Max Dex bonus. In Pathfinder they actually went the other way and gave Fighters the special ability to add more of their Dex bonus in heavy armour (which I appreciate even though the Pathfinder Fighter still sucks).
13th Age has an interesting solution to this whole thing in that it makes certain things based off the middle of three scores. Eg. AC is the middle score of Dex/Con/Wis, Physical Defence is Str/Con/Dex and Mental Defence is Int/Wis/Cha. In practice this still allows you to dump a score if you don't need it, and it still makes Con/Dex and Wis the more important scores, but it does offer some flexibility.
Part of the issue right now is that you have to consider the scaling of Ability Scores as well. Right now if I want to make a Str based Fighter in light armour (for some reason), I'm trading off AC for To hit and Damage. That Trade-off is not such a bad thing at 1st level. I might decide that 14 Dex is fine and 16 Str is good to hit. On balance this is fine, I'm one point down on AC compared to a light armour finesse using specialist. I could live with that at 1st level. The problem is that light armour assumes that I will be raising my Dex, however I can either raise my Strength and improve my chance to hit, or I Can raise my Dex, but I can't do both (or I can, but I'm still behind the curve on both). Introducing real trade-offs means rethinking the way that ability score bonuses are gained. (Personally I'd be tempted to go for +1 to 3 scores every four levels).
